\n
\n\n
\n\n
\n© 2024 Author(s). Published under an exclusive license by AIP Publishing.
\nPublished by Elsevier Ltd. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
\n\nSemantic mapping in 3D is fundamental to a wide range of geoscientific studies and applications, including geomorphology, hazard assessment, and environmental monitoring. However, automatically segmenting geological features from large-scale photogrammetric datasets remains a significant challenge. We present a methodology to address this gap. Using overlapping images collected over environments of interest, Structure-from-Motion (SfM) produces georeferenced point clouds and estimates camera poses. Existing large vision models, such as Segment Anything Model, segment objects in the images, generating pixel-segmentation associations. To produce pixel-point associations, we project the points back onto the camera image planes. As objects are independently segmented across multiple images with different perspectives, we develop a segmentation mosaicking algorithm to build probabilistic point-segmentation associations that combines the pixel-segmentation associations and pixel-point associations. Our methodology is validated using both synthetic data generated by Kubric and real-world UAV-SfM data. The implementation is designed to be compatible with existing SfM software, including Agisoft and OpenDroneMap, for photogrammetry mapping in geoscience studies. As a case study, we apply our method to the semantic mapping of precariously balanced rocks (PBRs), which provide upper-bound constraints on historical ground motion shaking intensity. To support object-level identification of PBRs, we additionally integrated Grounding DINO, enabling text-prompted segmentation of features of interest within UAV imagery. This case study demonstrates the effectiveness of our method in generating a 3D semantic map of PBRs, enabling spatial distribution of PBR fragility for earthquake hazard analysis.
", "date": "2026-07", "date_type": "published", "publication": "Computers & Geosciences", "volume": "213", "publisher": "Elsevier", "pagerange": "106181", "issn": "0098-3004", "official_url": "https://authors.library.caltech.edu/records/4tkjs-k7866", "funders": { "items": [ {}, {} ] }, "local_group": { "items": [ { "id": "Division-of-Geological-and-Planetary-Sciences" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" }, { "id": "Seismological-Laboratory" } ] }, "doi": "10.1016/j.cageo.2026.106181", "primary_object": { "basename": "1-s2.0-S0098300426000786-main.pdf", "url": "https://authors.library.caltech.edu/records/4tkjs-k7866/files/1-s2.0-S0098300426000786-main.pdf" }, "pub_year": "2026", "author_list": "Chen, Zhiang; McPhillips, Devin; et al." }, { "id": "https://authors.library.caltech.edu/records/61t51-jnn75", "eprint_status": "archive", "datestamp": "2025-06-05 18:39:12", "lastmod": "2025-06-05 18:39:12", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Gherold-Vincent", "name": { "family": "Gherold", "given": "Vincent" }, "orcid": "0009-0000-5870-5570" }, { "id": "Mandralis-Ioannis", "name": { "family": "Mandralis", "given": "Ioannis" }, "orcid": "0000-0001-5270-0672" }, { "id": "Sihite-Eric", "name": { "family": "Sihite", "given": "Eric" }, "orcid": "0000-0002-8653-8842" }, { "name": { "family": "Salagame", "given": "Adarsh" }, "orcid": "0000-0003-4345-3299" }, { "name": { "family": "Ramezani", "given": "Alireza" }, "orcid": "0000-0002-3391-5288" }, { "id": "Gharib-M", "name": { "family": "Gharib", "given": "Morteza (Mory)" }, "orcid": "0000-0003-0754-4193" } ] }, "title": "Self-supervised cost of transport estimation for multimodal path planning", "ispublished": "pub", "full_text_status": "public", "keywords": "AI-enabled robotics; field robots; motion and path planning; vision-based navigation", "note": "© 2025 IEEE. All rights reserved, including rights for text and data mining, and training of artificial intelligence and similar technologies.
\n\nThe authors would like to acknowledge funding from the Center for Autonomous Systems and Technology (CAST) at Caltech. IM is supported by the Onassis Foundation and the GALCIT graduate student endowment.
\n\nCode available at https://github.com/VinceGHER/self_supervised_cot_estimation.
\n\nThis article was recommended for publication by Associate Editor F. Rameau and Editor P. Vasseur upon evaluation of the reviewers’ comments.
", "abstract": "© 2024 Elsevier.
\n\nBH and HO acknowledge support from the Jet Propulsion Laboratory, California Institute of Technology , under a contract with the National Aeronautics and Space Administration and from Beyond Limits (Learning Optimal Models) through CAST (The Caltech Center for Autonomous Systems and Technologies).
\n\nBoumediene Hamzi: Conceptualization, Investigation, Methodology. Marcus Hutter: Conceptualization, Investigation, Methodology. Houman Owhadi: Conceptualization, Funding acquisition, Investigation, Methodology.
\n\nNo data was used for the research described in the article.
\n\nNo conflict of interest.
", "abstract": "Machine Learning (ML) and Algorithmic Information Theory (AIT) look at Complexity from different points of view. We explore the interface between AIT and Kernel Methods (that are prevalent in ML) by adopting an AIT perspective on the problem of learning kernels from data, in kernel ridge regression, through the method of Sparse Kernel Flows. In particular, by looking at the differences and commonalities between Minimal Description Length (MDL) and Regularization in Machine Learning (RML), we prove that the method of Sparse Kernel Flows is the natural approach to adopt to learn kernels from data. This approach aligns naturally with the MDL principle, offering a more robust theoretical basis than the existing reliance on cross-validation. The study reveals that deriving Sparse Kernel Flows does not require a statistical approach; instead, one can directly engage with code-lengths and complexities, concepts central to AIT. Thereby, this approach opens the door to reformulating algorithms in machine learning using tools from AIT, with the aim of providing them a more solid theoretical foundation.
\n© 2024 The Author(s). Published by Elsevier Under a Creative Commons license.
\n\nBH thanks Prof. Jeroen Lamb (Imperial College London) for useful discussions about random dynamical systems that inspired the work in this paper. BH acknowledge support from the Jet Propulsion Laboratory, USA, California Institute of Technology, USA , under a contract with the National Aeronautics and Space Administration and from Beyond Limits (Learning Optimal Models) through CAST (The Caltech Center for Autonomous Systems and Technologies). KD thanks Muhammad Alaskandarani for useful discussions and work on the early parts of this work. This work has been partially supported by the Gulf University for Science and Technology, including by project code: ISG Case 9.
\n\nBoumediene Hamzi: Methodology, Investigation, Conceptualization. Kamaludin Dingle: Methodology, Investigation, Conceptualization.
\n\nNo data was used for the research described in the article.
\n\nThe authors declare that the work presented in this paper is original and has not been published elsewhere in any form or language (partially or in full), except in abstract format for conferences. This work is the result of our research conducted primarily at the Gulf University for Science and Technology and has been partially supported by project code: ISG Case 9. We have no conflicts of interest to disclose and affirm that all sources used have been appropriately credited following ethical research practices. Furthermore, we agree to the terms of submission and publication in the journal to which we are submitting, and we confirm that all co-authors have approved the manuscript for submission.
", "abstract": "Simplicity bias is an intriguing phenomenon prevalent in various input–output maps, characterized by a preference for simpler, more regular, or symmetric outputs. Notably, these maps typically feature high-probability outputs with simple patterns, whereas complex patterns are exponentially less probable. This bias has been extensively examined and attributed to principles derived from algorithmic information theory and algorithmic probability. In a significant advancement, it has been demonstrated that the renowned logistic map \ud835\udc65\ud835\udc58+1=\ud835\udf07\ud835\udc65\ud835\udc58(1−\ud835\udc65\ud835\udc58), a staple in dynamical systems theory, and other one-dimensional maps exhibit simplicity bias when conceptualized as input–output systems. Building upon this work, our research delves into the manifestations of simplicity bias within the random logistic map, specifically focusing on scenarios involving additive noise. This investigation is driven by the overarching goal of formulating a comprehensive theory for the prediction and analysis of time series.
\nOur primary contributions are multifaceted. We discover that simplicity bias is observable in the random logistic map for specific ranges of \ud835\udf07 and noise magnitudes. Additionally, we find that this bias persists even with the introduction of small measurement noise, though it diminishes as noise levels increase. Our studies also revisit the phenomenon of noise-induced chaos, particularly when \ud835\udf07=3.83, revealing its characteristics through complexity-probability plots. Intriguingly, we employ the logistic map to illustrate a paradoxical aspect of data analysis: more data adhering to a consistent trend can occasionally lead to reduced confidence in extrapolation predictions, challenging conventional wisdom.
\nWe propose that adopting a probability-complexity perspective in analyzing dynamical systems could significantly enrich statistical learning theories related to series prediction and analysis. This approach not only facilitates a deeper understanding of simplicity bias and its implications but also paves the way for novel methodologies in forecasting complex systems behavior, especially in scenarios dominated by uncertainty and stochasticity.
", "date": "2024-07", "date_type": "published", "publication": "Physica D: Nonlinear Phenomena", "volume": "463", "publisher": "Elsevier", "pagerange": "134160", "issn": "0167-2789", "official_url": "https://authors.library.caltech.edu/records/98t71-esc27", "funders": { "items": [ {}, { "grant_number": "ISG Case 9" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1016/j.physd.2024.134160", "primary_object": { "basename": "1-s2.0-S0167278924001118-main.pdf", "url": "https://authors.library.caltech.edu/records/98t71-esc27/files/1-s2.0-S0167278924001118-main.pdf" }, "pub_year": "2024", "author_list": "Hamzi, Boumediene and Dingle, Kamaludin" }, { "id": "https://authors.library.caltech.edu/records/83c93-zb655", "eprint_status": "archive", "datestamp": "2024-05-02 19:37:08", "lastmod": "2024-05-02 19:37:08", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "O'Connell-Michael-T", "name": { "family": "O'Connell", "given": "Michael" }, "orcid": "0000-0001-6681-8823" }, { "id": "Cho-Joshua", "name": { "family": "Cho", "given": "Joshua" }, "orcid": "0009-0000-5482-0010" }, { "id": "Anderson-Matthew-J", "name": { "family": "Anderson", "given": "Matthew" }, "orcid": "0000-0001-8884-3448" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Learning-Based Minimally-Sensed Fault-Tolerant Adaptive Flight Control", "ispublished": "pub", "full_text_status": "public", "keywords": "Artificial Intelligence; Control and Optimization; Computer Science Applications; Computer Vision and Pattern Recognition; Mechanical Engineering; Human-Computer Interaction; Biomedical Engineering; Control and Systems Engineering", "note": "This work was supported in part by Supernal, LLC, and in part by Defense Advanced Research Projects Agency (DARPA). Video: https://youtu.be/IzFFEcvQiXw.
\n\n© 2024 IEEE.
", "abstract": "© 2024 IEEE.
\n\nThe NASA University Leadership Initiative (grant #80NSSC20M0163) provided funds to assist the authors with their research, but this article solely reflects the opinions and conclusions of its authors and not any NASA entity. Toyota Research Institute provided funds to support this work. The first author was supported by a NASA NSTGRO Fellowship, and the second author was supported on a NASA NSTRF Fellowship.
\nWe would like to thank Keiko Nagami, Adam Caccavale, Gadi Camps, and Jun En Low for their insights throughout this project.
\n\nOur code can be found at https://github.com/chengine/catnips.
", "abstract": "E.B. was supported by the the Foster and Coco Stanback Postdoctoral Fellowship. A.S. was supported by the Office of Naval Research (ONR) through Grant No. N00014-17-1-2079. T.C. and A.S. acknowledge recent support through ONR Grant No. N00014-23-1-2654. E.B. and A.S. are also grateful for support from the Department of Defense Vannevar Bush Faculty Fellowship held by A.S. We thank Tapio Schneider, Dimitris Giannakis, and two anonymous referees for helpful comments.
\nEviatar Bach: Conceptualization (equal); Formal analysis (equal); Investigation (lead); Methodology (equal); Software (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (equal). Tim Colonius: Conceptualization (supporting); Funding acquisition (equal); Supervision (supporting); Writing – review & editing (supporting). Isabel Scherl: Conceptualization (supporting); Investigation (supporting); Writing – review & editing (supporting). Andrew Stuart: Conceptualization (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal).
\nThe data that support the findings of this study are available within the article.
\nThe authors have no conflicts to disclose.
\nWe consider the problem of filtering dynamical systems, possibly stochastic, using observations of statistics. Thus, the computational task is to estimate a time-evolving density ρ(v,t) given noisy observations of the true density ρ†; this contrasts with the standard filtering problem based on observations of the state v. The task is naturally formulated as an infinite-dimensional filtering problem in the space of densities ρ. However, for the purposes of tractability, we seek algorithms in state space; specifically, we introduce a mean-field state-space model, and using interacting particle system approximations to this model, we propose an ensemble method. We refer to the resulting methodology as the ensemble Fokker–Planck filter (EnFPF). Under certain restrictive assumptions, we show that the EnFPF approximates the Kalman–Bucy filter for the Fokker–Planck equation, which is the exact solution to the infinite-dimensional filtering problem. Furthermore, our numerical experiments show that the methodology is useful beyond this restrictive setting. Specifically, the experiments show that the EnFPF is able to correct ensemble statistics, to accelerate convergence to the invariant density for autonomous systems, and to accelerate convergence to time-dependent invariant densities for non-autonomous systems. We discuss possible applications of the EnFPF to climate ensembles and to turbulence modeling.
", "date": "2024-03", "date_type": "published", "publication": "Chaos: An Interdisciplinary Journal of Nonlinear Science", "volume": "34", "number": "3", "publisher": "American Institute of Physics", "pagerange": "033119", "issn": "1054-1500", "official_url": "https://authors.library.caltech.edu/records/ce1pw-npb18", "funders": { "items": [ { "grant_number": "Foster and Coco Stanback Postdoctoral Fellowship" }, { "grant_number": "N00014-17-1-2079" }, { "grant_number": "N00014-23-1-2654" }, { "grant_number": "Vannevar Bush Faculty Fellowship" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1063/5.0171827", "primary_object": { "basename": "033119_1_5.0171827.pdf", "url": "https://authors.library.caltech.edu/records/ce1pw-npb18/files/033119_1_5.0171827.pdf" }, "pub_year": "2024", "author_list": "Bach, Eviatar; Colonius, Tim; et al." }, { "id": "https://authors.library.caltech.edu/records/adwww-nw726", "eprint_status": "archive", "datestamp": "2023-09-28 17:14:48", "lastmod": "2025-11-22 02:06:24", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Dabiri-J-O", "name": { "family": "Dabiri", "given": "John O." }, "orcid": "0000-0002-6722-9008" }, { "id": "Howland-Michael-F", "name": { "family": "Howland", "given": "Michael F." }, "orcid": "0000-0002-2878-3874" }, { "id": "Fu-Matthew-K", "name": { "family": "Fu", "given": "Matthew K." }, "orcid": "0000-0003-3949-7838" }, { "id": "Goldshmid-Roni-H", "name": { "family": "Goldshmid", "given": "Roni H." }, "orcid": "0000-0001-9095-3259" } ] }, "title": "Visual anemometry for physics-informed inference of wind", "ispublished": "pub", "full_text_status": "public", "keywords": "General Physics and Astronomy", "note": "\u00a9 Springer Nature Limited 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
\n\nThe authors gratefully acknowledge seminal contributions from J.L. Cardona in development of several of the concepts presented in this Perspective article, as well as discussions with K. Bouman, J. Sun, Y. Yue and P. Perona at Caltech. Additional helpful discussions occurred in the CV4Ecology Summer Workshop, supported by the Caltech Resnick Sustainability Institute. Constructive feedback from the anonymous reviewers led to meaningful improvements to the presentation of the material in this manuscript. Funding was generously provided by the National Science Foundation (Grant CBET-2019712) and the Center for Autonomous Systems and Technologies at Caltech. Additional support from Heliogen is gratefully acknowledged.
\n\nAll authors contributed to all aspects of the manuscript.
\n\nThe authors declare no competing interests.
", "abstract": "Accurate measurements of atmospheric flows at metre-scale resolution are essential for many sustainability applications, including optimal design of wind and solar farms, navigation and control of air flows in the built environment, monitoring of environmental phenomena such as wildfires and air pollution dispersal, and data assimilation into weather and climate models. Measurement of the relevant multiscale wind flows is inherently challenged by the optical transparency of the wind. This Perspective article explores new ways in which physics can be leveraged to 'see' environmental flows non-intrusively, that is, without the need to place measurement instruments directly in the flows of interest. Specifically, although wind itself is transparent, its effect can be seen in the motion of objects embedded in the environment and subjected to wind \u2014 swaying trees and flapping flags are commonly encountered examples. We survey emerging efforts to accomplish visual anemometry, the task of quantitatively inferring local wind conditions on the basis of the physics of observed flow\u2013structure interactions. Approaches based on first-principles physics as well as data-driven, machine learning methods will be described, and remaining obstacles to fully generalizable visual anemometry are discussed.
", "date": "2023-08-22", "date_type": "published", "publication": "Nature Reviews Physics", "publisher": "Nature Publishing Group", "issn": "2522-5820", "official_url": "https://authors.library.caltech.edu/records/adwww-nw726", "funders": { "items": [ { "grant_number": "CBET-2019712" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1038/s42254-023-00626-8", "pub_year": "2023", "author_list": "Dabiri, John O.; Howland, Michael F.; et al." }, { "id": "https://authors.library.caltech.edu/records/arcfd-g6b11", "eprint_status": "archive", "datestamp": "2023-11-10 18:08:58", "lastmod": "2025-11-22 03:10:34", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Sihite-Eric", "name": { "family": "Sihite", "given": "Eric" }, "orcid": "0000-0002-8653-8842" }, { "id": "Kalantari-Arash", "name": { "family": "Kalantari", "given": "Arash" }, "orcid": "0000-0003-2184-8548" }, { "id": "Nemovi-Reza", "name": { "family": "Nemovi", "given": "Reza" } }, { "id": "Ramezani-Alireza", "name": { "family": "Ramezani", "given": "Alireza" }, "orcid": "0000-0002-3391-5288" }, { "id": "Gharib-M", "name": { "family": "Gharib", "given": "Morteza (Mory)" }, "orcid": "0000-0003-0754-4193" } ] }, "title": "Publisher Correction: Multi-Modal Mobility Morphobot (M4) with appendage repurposing for locomotion plasticity enhancement", "ispublished": "pub", "full_text_status": "public", "keywords": "General Physics and Astronomy; General Biochemistry, Genetics and Molecular Biology; General Chemistry; Multidisciplinary", "note": "\u00a9 The Author(s) 2023. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
", "abstract": "Correction to: Nature Communications 10.1038/s41467-023-39018-y, published online 27 June 2023.
The original version of this Article omitted a reference to previous work in 'Meiri, N. et al. Flying STAR, a hybrid crawling and flying sprawl tuned robot, In 2019 IEEE International Conference on Robotics and Automation (ICRA), 5302\u20135308 (IEEEE, 2019).' This has been added as reference 44 in the sixth line of the seventh paragraph of the introduction: \"However, M4 differs from refs. 42, 43, 44 work because M4 exhaust appendage redundancy manipulation through morphing to maximize locomotion plasticity.\" This has been corrected in the PDF and HTML versions of the Article.
The Peer Review File associated with this Article was updated shortly after publication to remove the mistakenly copied article file.
", "date": "2023-08-07", "date_type": "published", "publication": "Nature Communications", "volume": "14", "publisher": "Nature Publishing Group", "pagerange": "4740", "issn": "2041-1723", "official_url": "https://authors.library.caltech.edu/records/arcfd-g6b11", "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1038/s41467-023-40466-9", "pmcid": "PMC10406874", "primary_object": { "basename": "s41467-023-40466-9.pdf", "url": "https://authors.library.caltech.edu/records/arcfd-g6b11/files/s41467-023-40466-9.pdf" }, "pub_year": "2023", "author_list": "Sihite, Eric; Kalantari, Arash; et al." }, { "id": "https://authors.library.caltech.edu/records/pch7d-bq204", "eprint_id": 122006, "eprint_status": "archive", "datestamp": "2023-08-22 21:15:47", "lastmod": "2026-03-28 00:22:04", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Sihite-Eric", "name": { "family": "Sihite", "given": "Eric" }, "orcid": "0000-0002-8653-8842" }, { "id": "Kalantari-Arash", "name": { "family": "Kalantari", "given": "Arash" }, "orcid": "0000-0003-2184-8548" }, { "id": "Nemovi-Reza", "name": { "family": "Nemovi", "given": "Reza" } }, { "id": "Ramezani-Alireza", "name": { "family": "Ramezani", "given": "Alireza" }, "orcid": "0000-0002-3391-5288" }, { "id": "Gharib-M", "name": { "family": "Gharib", "given": "Morteza" }, "orcid": "0000-0003-0754-4193" } ] }, "title": "Multi-Modal Mobility Morphobot (M4) with appendage repurposing for locomotion plasticity enhancement", "ispublished": "pub", "full_text_status": "public", "keywords": "General Physics and Astronomy; General Biochemistry, Genetics and Molecular Biology; General Chemistry; Multidisciplinary", "note": "\u00a9 The Author(s) 2023. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. \n\nThis project is funded by Caltech's Jet Propulsion Laboratory. A.R. efforts were partly funded by an NSF Foundational Research in Robotics (FRR), Award # 2142519, and a JPL Faculty Research Program (JFRP) fund. We acknowledge the work of graduate students and engineers at Caltech and Northeastern University for their help in prototyping and testing M4. Specifically, we are thankful to Dr. Milad Ramezani at Commonwealth Scientific and Industrial Research Organisation (CSIRO) and Filip Slezak from Swiss Federal Institute of Technology Lausanne (EPFL) for their help with autonomous multi-modal UAS-UGV operations. Noel Esparza-Duran supported the prototyping of M4. Benjamin Mottis from EPFL helped with preliminary path planning simulations and experiments. \n\nContributions. E.S. led the prototyping efforts, simulations, and experimentation. E.S. and A.R. collaboratively wrote the draft. A.K. evaluated the presented multi-modal models. R.N. supported prototyping efforts. A.K. and M.G. helped with draft editing. A.R. and M.G. conceived the M4 idea and are the principal investigators. \n\nData availability. Data will be provided upon request. \n\nCode availability. Simulation and path planning codes will be provided upon request. \n\nThe authors declare no competing interests.\n\nPublished - s41467-023-39018-y.pdf
Supplemental Material - 41467_2023_39018_MOESM10_ESM.mp4
Supplemental Material - 41467_2023_39018_MOESM1_ESM.pdf
Supplemental Material - 41467_2023_39018_MOESM3_ESM.pdf
Supplemental Material - 41467_2023_39018_MOESM4_ESM.mp4
Supplemental Material - 41467_2023_39018_MOESM5_ESM.mp4
Supplemental Material - 41467_2023_39018_MOESM6_ESM.mp4
Supplemental Material - 41467_2023_39018_MOESM7_ESM.mp4
Supplemental Material - 41467_2023_39018_MOESM8_ESM.mp4
Supplemental Material - 41467_2023_39018_MOESM9_ESM.mp4
", "abstract": "Robot designs can take many inspirations from nature, where there are many examples of highly resilient and fault-tolerant locomotion strategies to navigate complex terrains by recruiting multi-functional appendages. For example, birds such as Chukars and Hoatzins can repurpose wings for quadrupedal walking and wing-assisted incline running. These animals showcase impressive dexterity in employing the same appendages in different ways and generating multiple modes of locomotion, resulting in highly plastic locomotion traits which enable them to interact and navigate various environments and expand their habitat range. The robotic biomimicry of animals' appendage repurposing can yield mobile robots with unparalleled capabilities. Taking inspiration from animals, we have designed a robot capable of negotiating unstructured, multi-substrate environments, including land and air, by employing its components in different ways as wheels, thrusters, and legs. This robot is called the Multi-Modal Mobility Morphobot, or M4 in short. M4 can employ its multi-functional components composed of several actuator types to (1) fly, (2) roll, (3) crawl, (4) crouch, (5) balance, (6) tumble, (7) scout, and (8) loco-manipulate. M4 can traverse steep slopes of up to 45 deg. and rough terrains with large obstacles when in balancing mode. M4 possesses onboard computers and sensors and can autonomously employ its modes to negotiate an unstructured environment. We present the design of M4 and several experiments showcasing its multi-modal capabilities.", "date": "2023-06-27", "date_type": "published", "publication": "Nature Communications", "volume": "14", "publisher": "Nature Publishing Group", "pagerange": "Art. No. 3323", "id_number": "CaltechAUTHORS:20230628-295324000.2", "issn": "2041-1723", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20230628-295324000.2", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "JPL/Caltech" }, { "agency": "NSF", "grant_number": "CMMI-2142519" }, { "agency": "JPL Faculty Research Program" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" }, { "id": "GALCIT" } ] }, "doi": "10.1038/s41467-023-39018-y", "pmcid": "PMC10300070", "primary_object": { "basename": "s41467-023-39018-y.pdf", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/s41467-023-39018-y.pdf" }, "related_objects": [ { "basename": "41467_2023_39018_MOESM9_ESM.mp4", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM9_ESM.mp4" }, { "basename": "41467_2023_39018_MOESM1_ESM.pdf", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM1_ESM.pdf" }, { "basename": "41467_2023_39018_MOESM5_ESM.mp4", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM5_ESM.mp4" }, { "basename": "41467_2023_39018_MOESM7_ESM.mp4", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM7_ESM.mp4" }, { "basename": "41467_2023_39018_MOESM10_ESM.mp4", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM10_ESM.mp4" }, { "basename": "41467_2023_39018_MOESM3_ESM.pdf", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM3_ESM.pdf" }, { "basename": "41467_2023_39018_MOESM4_ESM.mp4", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM4_ESM.mp4" }, { "basename": "41467_2023_39018_MOESM6_ESM.mp4", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM6_ESM.mp4" }, { "basename": "41467_2023_39018_MOESM8_ESM.mp4", "url": "https://authors.library.caltech.edu/records/pch7d-bq204/files/41467_2023_39018_MOESM8_ESM.mp4" } ], "pub_year": "2023", "author_list": "Sihite, Eric; Kalantari, Arash; et al." }, { "id": "https://authors.library.caltech.edu/records/qpmpg-p1h18", "eprint_id": 121632, "eprint_status": "archive", "datestamp": "2023-08-22 20:53:41", "lastmod": "2025-11-22 03:58:23", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Tang-Ellande", "name": { "family": "Tang", "given": "Ellande" }, "orcid": "0000-0001-5933-4716" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Experiments and Modeling of the Ceiling Effect with Drone-Scale Propellers", "ispublished": "pub", "full_text_status": "public", "keywords": "Propeller Performance, Uninhabited Aerial Vehicle, Disk Actuator Theory, Computational Fluid Dynamics, VTOL Aircraft Design, Free Stream Velocity, Rotorcrafts, Vortex Filaments, Finite Element Software, Differential Pressure Sensors; Aerospace Engineering", "note": "\u00a9 2023 by the American Institute of Aeronautics and Astronautics, Inc. \n\nThis work was supported by a National Defense Science and Engineering Graduate fellowship administered via the U.S. Air Force Office of Scientific Research. Experimental setups used resources provided by the Center for Autonomous Systems and Technology at the California Institute of Technology. The authors thank M. Anderson for his constructive comments.", "abstract": "Aircraft designs involving aerodynamic interactions between rotors and other aerodynamic surfaces are becoming increasingly popular. Studying these interactions is key to developing good models for design and aspects of the system such as control. This paper seeks to understand the effect of obstructing the upstream of a propeller at varying distances, also called the ceiling effect, by examining in depth the canonical case of a theoretically infinite obstruction and its interaction with a propeller or actuator disk. Experimental studies of force and pressure are compared to results from computational fluid dynamics (CFD), and good agreement is shown. The CFD results are then compared against the Morillo flowfield model, and a correction is found to match the results. The data indicate that some propellers experienced a nearly twofold increase in thrust. However, a matching force on the surface develops, and the net force drops to nearly zero. The force interaction between the two nearly disappears once the separation exceeds half a propeller diameter. These results are independent of propeller size and pitch. The implemented theoretical model also has low computational cost, and it could be used to improve low-order models such as panel methods or provide a foundation for future rotor\u2013body interaction modeling.", "date": "2023-06-01", "date_type": "published", "publication": "AIAA Journal", "publisher": "AIAA", "id_number": "CaltechAUTHORS:20230530-441768000.72", "issn": "0001-1452", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20230530-441768000.72", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Center for Autonomous Systems and Technologies" }, { "agency": "National Defense Science and Engineering Graduate (NDSEG) Fellowship" }, { "agency": "Bren Professor of Control and Dynamical Systems" } ] }, "other_numbering_system": { "items": [ { "id": "2021-1648", "name": "AIAA Paper" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.2514/1.j062568", "pub_year": "2023", "author_list": "Tang, Ellande and Chung, Soon-Jo" }, { "id": "https://authors.library.caltech.edu/records/6ygz3-0b238", "eprint_id": 122116, "eprint_status": "archive", "datestamp": "2023-08-22 20:50:19", "lastmod": "2026-03-27 18:24:47", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Dorobantu-Victor-D", "name": { "family": "Dorobantu", "given": "Victor D." }, "orcid": "0000-0002-2797-7802" }, { "id": "Azizzadenesheli-Kamyar", "name": { "family": "Azizzadenesheli", "given": "Kamyar" }, "orcid": "0000-0001-8507-1868" }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" } ] }, "title": "Compactly Restrictable Metric Policy Optimization Problems", "ispublished": "pub", "full_text_status": "public", "keywords": "Electrical and Electronic Engineering; Computer Science Applications; Control and Systems Engineering", "note": "\u00a9 2022 IEEE. \n\nThis work was supported in part by DARPA and in part by Beyond Limits. The work of Victor D. Dorobantu was supported by a Kortschak Fellowship.", "abstract": "We study policy optimization problems for deterministic Markov decision processes (MDPs) with metric state and action spaces, which we refer to as metric policy optimization problems (MPOPs). Our goal is to establish theoretical results on the well-posedness of MPOPs that can characterize practically relevant continuous control systems. To do so, we define a special class of MPOPs called compactly restrictable MPOPs (CR-MPOPs), which are flexible enough to capture the complex behavior of robotic systems but specific enough to admit solutions using dynamic programming methods such as value iteration. We show how to arrive at CR-MPOPs using forward-invariance. We further show that our theoretical results on CR-MPOPs can be used to characterize feedback linearizable control affine systems.", "date": "2023-05", "date_type": "published", "publication": "IEEE Transactions on Automatic Control", "volume": "68", "number": "5", "publisher": "IEEE", "pagerange": "3115-3122", "id_number": "CaltechAUTHORS:20230705-704041500.16", "issn": "0018-9286", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20230705-704041500.16", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Defense Advanced Research Projects Agency (DARPA)" }, { "agency": "Beyond Limits" }, { "agency": "Kortschak Scholars Program" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/tac.2022.3217269", "pub_year": "2023", "author_list": "Dorobantu, Victor D.; Azizzadenesheli, Kamyar; et al." }, { "id": "https://authors.library.caltech.edu/records/mp0mq-5z817", "eprint_id": 119820, "eprint_status": "archive", "datestamp": "2023-08-22 18:53:26", "lastmod": "2026-03-27 18:29:46", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Darcy-Matthieu", "name": { "family": "Darcy", "given": "Matthieu" }, "orcid": "0000-0003-0029-091X" }, { "id": "Hamzi-Boumediene", "name": { "family": "Hamzi", "given": "Boumediene" }, "orcid": "0000-0002-9446-2614" }, { "id": "Livieri-Giulia", "name": { "family": "Livieri", "given": "Giulia" }, "orcid": "0000-0002-3777-7329" }, { "id": "Owhadi-H", "name": { "family": "Owhadi", "given": "Houman" }, "orcid": "0000-0002-5677-1600" }, { "id": "Tavallali-Peyman", "name": { "family": "Tavallali", "given": "Peyman" }, "orcid": "0000-0001-7166-5489" } ] }, "title": "One-shot learning of stochastic differential equations with data adapted kernels", "ispublished": "pub", "full_text_status": "public", "keywords": "Condensed Matter Physics; Statistical and Nonlinear Physics", "note": "\u00a9 2022 Elsevier. \n\nMD, BH, HO acknowledge partial support by the Air Force Office of Scientific Research, USA under MURI award number FA9550-20-1-0358 (Machine Learning and Physics-Based Modeling and Simulation). MD, PT and HO acknowledge support from Beyond Limits (Learning Optimal Models) through CAST (The Caltech Center for Autonomous Systems and Technologies).", "abstract": "We consider the problem of learning Stochastic Differential Equations of the form dX\u209c = f (X\u209c)d\u209c + \u03c3(X\u209c)dW\u209c from one sample trajectory. This problem is more challenging than learning deterministic dynamical systems because one sample trajectory only provides indirect information on the unknown functions f, \u03c3, and stochastic process dW\u209c representing the drift, the diffusion, and the stochastic forcing terms, respectively. We propose a method that combines Computational Graph Completion [1] and data adapted kernels learned via a new variant of cross validation. Our approach can be decomposed as follows: (1) Represent the time-increment map X\u209c \u2192 X_(t+dt) as a Computational Graph in which f, \u03c3 and dW\u209c appear as unknown functions and random variables. (2) Complete the graph (approximate unknown functions and random variables) via Maximum a Posteriori Estimation (given the data) with Gaussian Process (GP) priors on the unknown functions. (3) Learn the covariance functions (kernels) of the GP priors from data with randomized cross-validation. Numerical experiments illustrate the efficacy, robustness, and scope of our method.", "date": "2023-02", "date_type": "published", "publication": "Physica D", "volume": "444", "publisher": "Elsevier", "pagerange": "Art. No. 133583", "id_number": "CaltechAUTHORS:20230307-205876300.8", "issn": "0167-2789", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20230307-205876300.8", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Air Force Office of Scientific Research (AFOSR)", "grant_number": "FA9550-20-1-0358" }, { "agency": "Beyond Limits" }, { "agency": "Center for Autonomous Systems and Technologies" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1016/j.physd.2022.133583", "pub_year": "2023", "author_list": "Darcy, Matthieu; Hamzi, Boumediene; et al." }, { "id": "https://authors.library.caltech.edu/records/g57e8-z8965", "eprint_id": 114771, "eprint_status": "archive", "datestamp": "2023-08-22 16:27:55", "lastmod": "2025-11-21 02:43:16", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Nakka-Yashwanth-Kumar-K", "name": { "family": "Nakka", "given": "Yashwanth Kumar K." }, "orcid": "0000-0001-7897-3644" }, { "id": "H\u00f6nig-Wolfgang", "name": { "family": "H\u00f6nig", "given": "Wolfgang" }, "orcid": "0000-0002-0773-028X" }, { "id": "Choi-Changrak", "name": { "family": "Choi", "given": "Changrak" } }, { "id": "Harvard-Alexei", "name": { "family": "Harvard", "given": "Alexei" } }, { "id": "Rahmani-Amir", "name": { "family": "Rahmani", "given": "Amir" } }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Information-Based Guidance and Control Architecture for Multi-Spacecraft On-Orbit Inspection", "ispublished": "pub", "full_text_status": "public", "keywords": "Applied Mathematics; Electrical and Electronic Engineering; Space and Planetary Science; Aerospace Engineering; Control and Systems Engineering", "note": "\u00a9 2022 by The Authors. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission. \n\nReceived 17 June 2021. Accepted 27 March 2022. Published online 15 May 2022. \n\nPart of this work was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract with NASA. This work was in part funded by the JPL-CAST Swarm Autonomy project and the David and Catherine Thompson Graduate Fellowship Fund for Space. The authors thank Fred Y. Hadaegh for technical discussions.", "abstract": "Inspection or mapping of a target spacecraft in a low Earth orbit using multiple observer spacecraft in stable passive relative orbits (PROs) is a key enabling technology for future space missions. Our guidance and control architecture uses an information gain approach to directly consider the tradeoff between gathered data and fuel/energy cost. The architecture has four components: information estimation, spacecraft's absolute and relative state estimation, motion planning for relative orbit initialization and reconfiguration, and relative orbit control. The information estimation quantifies the information gain during inspection of a spacecraft, given past and potential future poses of all spacecraft. The estimated information gain is a crucial input to the motion planner, which computes PROs and reconfiguration strategies for each observer to maximize the information gain from distributed observations of the target spacecraft. The resulting motion trajectories jointly consider observational coverage of the target spacecraft and fuel/energy cost. For the PRO trajectories, a fuel-optimal attitude trajectory that minimizes rest-to-rest energy for each observer to inspect the target spacecraft is designed. The validation on a mission simulation to visually inspect the target spacecraft and on a three-degree-of-freedom robotic spacecraft dynamics simulator testbed demonstrates the effectiveness and versatility of our approach.", "date": "2022-07", "date_type": "published", "publication": "Journal of Guidance, Control, and Dynamics", "volume": "45", "number": "7", "publisher": "AIAA", "pagerange": "1184-1201", "id_number": "CaltechAUTHORS:20220517-496840000", "issn": "0731-5090", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220517-496840000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "NASA/JPL/Caltech" }, { "agency": "David and Catherine Thompson Graduate Fellowship Fund for Space" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.2514/1.g006278", "pub_year": "2022", "author_list": "Nakka, Yashwanth Kumar K.; H\u00f6nig, Wolfgang; et al." }, { "id": "https://authors.library.caltech.edu/records/1mttm-4w433", "eprint_id": 113762, "eprint_status": "archive", "datestamp": "2023-08-22 16:26:06", "lastmod": "2025-11-20 23:42:34", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Chen-Xin", "name": { "family": "Chen", "given": "Xin" }, "orcid": "0000-0002-0952-0008" }, { "id": "Qu-Guannan", "name": { "family": "Qu", "given": "Guannan" }, "orcid": "0000-0002-5466-3550" }, { "id": "Tang-Yujie", "name": { "family": "Tang", "given": "Yujie" }, "orcid": "0000-0002-4921-8372" }, { "id": "Low-S-H", "name": { "family": "Low", "given": "Steven" }, "orcid": "0000-0001-6476-3048" }, { "id": "Li-Na", "name": { "family": "Li", "given": "Na" }, "orcid": "0000-0001-9545-3050" } ] }, "title": "Reinforcement Learning for Selective Key Applications in Power Systems: Recent Advances and Future Challenges", "ispublished": "pub", "full_text_status": "public", "keywords": "Frequency regulation, voltage control, energy management, reinforcement learning, smart grid; General Computer Science", "note": "\u00a9 2022 IEEE. \n\nManuscript received February 9, 2021; revised July 27, 2021 and November 17, 2021; accepted February 18, 2022. Date of publication February 25, 2022; date of current version June 21, 2022. \n\nThis work was supported in part by NSF CAREER under Grant ECCS-1553407; in part by the NSF AI Institute under Grant 2112085; in part by NSF under Grant ECCS-1931662, Grant AitF- 1637598, and Grant CNS-1518941; in part by Cyber-Physical Systems (CPS) under Grant ECCS-1932611; in part by Resnick Sustainability Institute; in part by PIMCO Fellowship; in part by Amazon AI4Science Fellowship; and in part by the Caltech Center for Autonomous Systems and Technologies (CAST). Paper no. TSG-00195-2021.\n\nAccepted Version - 2102.01168.pdf
", "abstract": "With large-scale integration of renewable generation and distributed energy resources, modern power systems are confronted with new operational challenges, such as growing complexity, increasing uncertainty, and aggravating volatility. Meanwhile, more and more data are becoming available owing to the widespread deployment of smart meters, smart sensors, and upgraded communication networks. As a result, data-driven control techniques, especially reinforcement learning (RL), have attracted surging attention in recent years. This paper provides a comprehensive review of various RL techniques and how they can be applied to decision-making and control in power systems. In particular, we select three key applications, i.e., frequency regulation, voltage control, and energy management, as examples to illustrate RL-based models and solutions. We then present the critical issues in the application of RL, i.e., safety, robustness, scalability, and data. Several potential future directions are discussed as well.", "date": "2022-07", "date_type": "published", "publication": "IEEE Transactions on Smart Grid", "volume": "13", "number": "4", "publisher": "IEEE", "pagerange": "2935-2958", "id_number": "CaltechAUTHORS:20220307-188369000", "issn": "1949-3053", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220307-188369000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "NSF", "grant_number": "ECCS-1553407" }, { "agency": "NSF", "grant_number": "CBET-2112085" }, { "agency": "NSF", "grant_number": "ECCS-1931662" }, { "agency": "NSF", "grant_number": "CCF-1637598" }, { "agency": "NSF", "grant_number": "CNS-1518941" }, { "agency": "NSF", "grant_number": "ECCS-1932611" }, { "agency": "Resnick Sustainability Institute" }, { "agency": "PIMCO" }, { "agency": "Amazon AI4Science Fellowship" }, { "agency": "Center for Autonomous Systems and Technologies" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" }, { "id": "Resnick-Sustainability-Institute" } ] }, "doi": "10.1109/tsg.2022.3154718", "primary_object": { "basename": "2102.01168.pdf", "url": "https://authors.library.caltech.edu/records/1mttm-4w433/files/2102.01168.pdf" }, "pub_year": "2022", "author_list": "Chen, Xin; Qu, Guannan; et al." }, { "id": "https://authors.library.caltech.edu/records/8ckmw-aj584", "eprint_id": 115363, "eprint_status": "archive", "datestamp": "2023-08-22 16:25:07", "lastmod": "2025-11-21 01:35:33", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Singletary-Andrew-W", "name": { "family": "Singletary", "given": "Andrew" }, "orcid": "0000-0001-6635-4256" }, { "id": "Ahmadi-Mohamadreza", "name": { "family": "Ahmadi", "given": "Mohamadreza" }, "orcid": "0000-0003-1447-3012" }, { "id": "Ames-A-D", "name": { "family": "Ames", "given": "Aaron D." }, "orcid": "0000-0003-0848-3177" } ] }, "title": "Safe Control for Nonlinear Systems With Stochastic Uncertainty via Risk Control Barrier Functions", "ispublished": "pub", "full_text_status": "public", "keywords": "Uncertain systems; Stochastic systems; Control and Optimization; Control and Systems Engineering", "note": "\u00a9 2022 IEEE. \n\nManuscript received 21 March 2022; revised 20 May 2022; accepted 11 June 2022. Date of publication 30 June 2022; date of current version 18 July 2022. \n\nThis work was supported in part by AeroVironment and\nin part by NSF CPS under Award 1932091. Recommended by Senior\nEditor L. Zhang\n\nAccepted Version - Safe_Control_for_Nonlinear_Systems_with_Stochastic_Uncertainty_via_Risk_Control_Barrier_Functions.pdf
Submitted - 2203.15892.pdf
", "abstract": "Guaranteeing safety for robotic and autonomous systems in real-world environments is a challenging task that requires the mitigation of stochastic uncertainties. Control barrier functions have, in recent years, been widely used for enforcing safety related set-theoretic properties, such as forward invariance and reachability, of nonlinear dynamical systems. In this letter, we extend this rich framework to nonlinear discrete-time systems subject to stochastic uncertainty and propose a framework for assuring risk-sensitive safety in terms of coherent risk measures. To this end, we introduce risk control barrier functions (RCBFs), which are compositions of barrier functions and dynamic, coherent risk measures. We show that the existence of such barrier functions implies invariance in a coherent risk sense. Furthermore, we formulate conditions based on finite-time RCBFs to guarantee finite-time reachability to a desired set in the coherent risk. Conditions for risk-sensitive safety and finite-time reachability of sets composed of Boolean compositions of multiple RCBF are also formulated. We show the efficacy of the proposed method through its application to a cart-pole system in a safety-critical scenario.", "date": "2022-06-30", "date_type": "published", "publication": "IEEE Control Systems Letters", "volume": "7", "publisher": "IEEE", "pagerange": "349-354", "id_number": "CaltechAUTHORS:20220707-315696000", "issn": "2475-1456", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220707-315696000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "AeroVironment" }, { "agency": "NSF", "grant_number": "CNS-1932091" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/lcsys.2022.3187458", "primary_object": { "basename": "2203.15892.pdf", "url": "https://authors.library.caltech.edu/records/8ckmw-aj584/files/2203.15892.pdf" }, "related_objects": [ { "basename": "Safe_Control_for_Nonlinear_Systems_with_Stochastic_Uncertainty_via_Risk_Control_Barrier_Functions.pdf", "url": "https://authors.library.caltech.edu/records/8ckmw-aj584/files/Safe_Control_for_Nonlinear_Systems_with_Stochastic_Uncertainty_via_Risk_Control_Barrier_Functions.pdf" } ], "pub_year": "2022", "author_list": "Singletary, Andrew; Ahmadi, Mohamadreza; et al." }, { "id": "https://authors.library.caltech.edu/records/q3grb-3vz72", "eprint_id": 114603, "eprint_status": "archive", "datestamp": "2023-08-22 15:30:09", "lastmod": "2025-11-20 20:16:46", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "O'Connell-Michael", "name": { "family": "O'Connell", "given": "Michael" }, "orcid": "0000-0001-6681-8823" }, { "id": "Shi-Guanya", "name": { "family": "Shi", "given": "Guanya" }, "orcid": "0000-0002-9075-3705" }, { "id": "Shi-Xichen", "name": { "family": "Shi", "given": "Xichen" }, "orcid": "0000-0002-5366-9256" }, { "id": "Azizzadenesheli-Kamyar", "name": { "family": "Azizzadenesheli", "given": "Kamyar" }, "orcid": "0000-0001-8507-1868" }, { "id": "Anandkumar-A", "name": { "family": "Anandkumar", "given": "Anima" }, "orcid": "0000-0002-6974-6797" }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Neural-Fly enables rapid learning for agile flight in strong winds", "ispublished": "pub", "full_text_status": "public", "keywords": "Artificial Intelligence; Control and Optimization; Computer Science Applications; Mechanical Engineering", "note": "\u00a9 2022 The Authors, some rights reserved; exclusive licensee\nAmerican Association for the Advancement of Science. No claim to original U.S. Government Works. \n\nSubmitted 11 October 2021; Accepted 12 April 2022; Published 4 May 2022. \n\nA.A. is also affiliated with NVIDIA Corporation, and Y.Y. is also with associated Argo AI. K.A. is currently affiliated with Purdue University. We thank J. Burdick and J.-J. E. Slotine for their helpful discussions. We thank M. Anderson for help with configuring the quadrotor platform, and M. Anderson and P. Spieler for help with hardware troubleshooting. We also thank N. Badillo and L. Pabon Madrid for help in experiments. \n\nThis research was developed with funding from the Defense Advanced Research Projects Agency (DARPA). This research was also conducted in part with funding from Raytheon Technologies. The views, opinions, and/or findings expressed are those of the authors and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government. The experiments reported in this article were conducted at Caltech's Center for Autonomous Systems and Technologies (CAST). \n\nAuthor contributions: S.-J.C. and Y.Y. directed the research activities. G.S. and M.O. designed and implemented the metalearning algorithm under the guidance of Y.Y., K.A., A.A., and S.-J.C., while the last-layer adaptation idea was started with a discussion by G.S., M.O., X.S., and S.-J.C. M.O. and G.S. designed and implemented the adaptive control algorithm with inputs from S.-J.C. and X.S. M.O. and G.S. performed experiments and evaluated the results. M.O. conducted the theoretical analysis of the meta-learning based adaptive controller with input from S.-J.C., G.S., and X.S. G.S. analyzed the learning algorithm with feedback from Y.Y., K.A., A.A., and S.-J.C. G.S. and M.O. created all the figures and videos with input from the other authors. All authors prepared the manuscript. \n\nThe authors declare that they have no competing interests. \n\nData and materials availability: All data needed to evaluate the conclusions in the article are present in the article or in the Supplementary Materials. We have provided the machine learning model training code, training data, and experimental data at github.com/aerorobotics/neural-fly.\n\nAccepted Version - 2205.06908.pdf
Supplemental Material - scirobotics.abm6597_sm.pdf
", "abstract": "Executing safe and precise flight maneuvers in dynamic high-speed winds is important for the ongoing commoditization of uninhabited aerial vehicles (UAVs). However, because the relationship between various wind conditions and its effect on aircraft maneuverability is not well understood, it is challenging to design effective robot controllers using traditional control design methods. We present Neural-Fly, a learning-based approach that allows rapid online adaptation by incorporating pretrained representations through deep learning. Neural-Fly builds on two key observations that aerodynamics in different wind conditions share a common representation and that the wind-specific part lies in a low-dimensional space. To that end, Neural-Fly uses a proposed learning algorithm, domain adversarially invariant meta-learning (DAIML), to learn the shared representation, only using 12 minutes of flight data. With the learned representation as a basis, Neural-Fly then uses a composite adaptation law to update a set of linear coefficients for mixing the basis elements. When evaluated under challenging wind conditions generated with the Caltech Real Weather Wind Tunnel, with wind speeds up to 43.6 kilometers/hour (12.1 meters/second), Neural-Fly achieves precise flight control with substantially smaller tracking error than stateof-the-art nonlinear and adaptive controllers. In addition to strong empirical performance, the exponential stability of Neural-Fly results in robustness guarantees. Last, our control design extrapolates to unseen wind conditions, is shown to be effective for outdoor flights with only onboard sensors, and can transfer across drones with minimal performance degradation.", "date": "2022-05-04", "date_type": "published", "publication": "Science Robotics", "volume": "7", "number": "66", "publisher": "American Association for the Advancement of Science", "pagerange": "Art. No. eabm6597", "id_number": "CaltechAUTHORS:20220505-792409800", "issn": "2470-9476", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220505-792409800", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Defense Advanced Research Projects Agency (DARPA)" }, { "agency": "Raytheon Company" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" }, { "id": "GALCIT" } ] }, "doi": "10.1126/scirobotics.abm6597", "primary_object": { "basename": "2205.06908.pdf", "url": "https://authors.library.caltech.edu/records/q3grb-3vz72/files/2205.06908.pdf" }, "related_objects": [ { "basename": "scirobotics.abm6597_sm.pdf", "url": "https://authors.library.caltech.edu/records/q3grb-3vz72/files/scirobotics.abm6597_sm.pdf" } ], "pub_year": "2022", "author_list": "O'Connell, Michael; Shi, Guanya; et al." }, { "id": "https://authors.library.caltech.edu/records/0m99v-t3788", "eprint_id": 114256, "eprint_status": "archive", "datestamp": "2023-08-22 15:21:48", "lastmod": "2025-11-21 04:05:02", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Wen-Gege", "name": { "family": "Wen", "given": "Gege" }, "orcid": "0000-0003-1668-3777" }, { "id": "Li-Zongyi", "name": { "family": "Li", "given": "Zongyi" }, "orcid": "0000-0003-2081-9665" }, { "id": "Azizzadenesheli-Kamyar", "name": { "family": "Azizzadenesheli", "given": "Kamyar" }, "orcid": "0000-0001-8507-1868" }, { "id": "Anandkumar-A", "name": { "family": "Anandkumar", "given": "Anima" }, "orcid": "0000-0002-6974-6797" }, { "id": "Benson-Sally-M", "name": { "family": "Benson", "given": "Sally M." }, "orcid": "0000-0002-3733-4296" } ] }, "title": "U-FNO\u2014An enhanced Fourier neural operator-based deep-learning model for multiphase flow", "ispublished": "pub", "full_text_status": "public", "keywords": "Multiphase flow; Fourier neural operator; Convolutional neural network; Carbon capture and storage; Deep learning; Water Science and Technology", "note": "\u00a9 2022 Elsevier. \n\nReceived 30 August 2021, Revised 7 February 2022, Accepted 25 March 2022, Available online 5 April 2022. \n\nG. Wen and S. M. Benson gratefully acknowledges the supported by ExxonMobil through the Strategic Energy Alliance at Stanford University and the Stanford Center for Carbon Storage . Z. Li gratefully acknowledges the financial support from the Kortschak Scholars Program. A. Anandkumar is supported in part by Bren endowed chair, LwLL grants, Beyond Limits, Raytheon, Microsoft, Google, Adobe faculty fellowships, and DE Logi grant. The authors would like to acknowledge the reviewers and editors for the constructive comments. \n\nCode and data availability: The python code for U-FNO model architecture and the data set used in training is available at https://github.com/gegewen/ufno. Web application https://ccsnet.ai hosts the trained U-FNO models to provide real time predictions. \n\nCRediT authorship contribution statement: Gege Wen: Conceptualization, Methodology, Software, Data curation, Formal analysis, Investigation, Validation, Visualization, Writing \u2013 original draft, Writing \u2013 review & editing. Zongyi Li: Conceptualization, Methodology, Software, Investigation, Validation, Writing \u2013 review & editing. Kamyar Azizzadenesheli: Methodology, Software, Investigation, Validation, Writing \u2013 review & editing. Anima Anandkumar: Funding acquisition, Supervision, Writing \u2013 review & editing. Sally M. Benson: Conceptualization, Formal analysis, Funding acquisition, Methodology, Resources, Supervision, Writing \u2013 review & editing. \n\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\n\nSubmitted - 2109.03697.pdf
", "abstract": "Numerical simulation of multiphase flow in porous media is essential for many geoscience applications. Machine learning models trained with numerical simulation data can provide a faster alternative to traditional simulators. Here we present U-FNO, a novel neural network architecture for solving multiphase flow problems with superior accuracy, speed, and data efficiency. U-FNO is designed based on the newly proposed Fourier neural operator (FNO), which has shown excellent performance in single-phase flows. We extend the FNO-based architecture to a highly complex CO\u2082-water multiphase problem with wide ranges of permeability and porosity heterogeneity, anisotropy, reservoir conditions, injection configurations, flow rates, and multiphase flow properties. The U-FNO architecture is more accurate in gas saturation and pressure buildup predictions than the original FNO and a state-of-the-art convolutional neural network (CNN) benchmark. Meanwhile, it has superior data utilization efficiency, requiring only a third of the training data to achieve the equivalent accuracy as CNN. U-FNO provides superior performance in highly heterogeneous geological formations and critically important applications such as gas saturation and pressure buildup \"fronts\" determination. The trained model can serve as a general-purpose alternative to routine numerical simulations of 2D-radial CO\u2082 injection problems with significant speed-ups than traditional simulators.", "date": "2022-05", "date_type": "published", "publication": "Advances in Water Resources", "volume": "163", "publisher": "Elsevier", "pagerange": "Art. No. 104180", "id_number": "CaltechAUTHORS:20220412-15492000", "issn": "0309-1708", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220412-15492000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "ExxonMobil Research and Engineering Company" }, { "agency": "Kortschak Scholars Program" }, { "agency": "Bren Professor of Computing and Mathematical Sciences" }, { "agency": "Learning with Less Labels (LwLL)" }, { "agency": "Beyond Limits" }, { "agency": "Raytheon Company" }, { "agency": "Microsoft Faculty Fellowship" }, { "agency": "Google Faculty Research Award" }, { "agency": "Adobe" }, { "agency": "Caltech De Logi Fund" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1016/j.advwatres.2022.104180", "primary_object": { "basename": "2109.03697.pdf", "url": "https://authors.library.caltech.edu/records/0m99v-t3788/files/2109.03697.pdf" }, "pub_year": "2022", "author_list": "Wen, Gege; Li, Zongyi; et al." }, { "id": "https://authors.library.caltech.edu/records/psxc9-vps89", "eprint_id": 107618, "eprint_status": "archive", "datestamp": "2023-08-20 07:23:23", "lastmod": "2025-11-21 03:34:44", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Shi-Guanya", "name": { "family": "Shi", "given": "Guanya" }, "orcid": "0000-0002-9075-3705" }, { "id": "H\u00f6nig-Wolfgang", "name": { "family": "H\u00f6nig", "given": "Wolfgang" }, "orcid": "0000-0002-0773-028X" }, { "id": "Shi-Xichen", "name": { "family": "Shi", "given": "Xichen" }, "orcid": "0000-0002-5366-9256" }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Neural-Swarm2: Planning and Control of Heterogeneous Multirotor Swarms Using Learned Interactions", "ispublished": "pub", "full_text_status": "public", "keywords": "Aerial systems, deep learning in robotics, multirobot motion planning and control, multirobot systems", "note": "\u00a9 2021 IEEE. \n\nManuscript received November 24, 2020; revised May 16, 2021; accepted June 15, 2021. Date of publication August 6, 2021; date of current version April 5, 2022. \n\nThe work was supported in part by Caltech's Center for Autonomous Systems and Technologies (CAST), the Raytheon Company, and the Jet Propulsion Laboratory. This paper was recommended for publication by Associate Editor M. Schwager and Editor P. R. Giordano upon evaluation of the reviewers' comments.\n\nSubmitted - 2012.05457.pdf
", "abstract": "We present Neural-Swarm2 , a learning-based method for motion planning and control that allows heterogeneous multirotors in a swarm to safely fly in close proximity. Such operation for drones is challenging due to complex aerodynamic interaction forces, such as downwash generated by nearby drones and ground effect. Conventional planning and control methods neglect capturing these interaction forces, resulting in sparse swarm configuration during flight. Our approach combines a physics-based nominal dynamics model with learned deep neural networks with strong Lipschitz properties. We make use of two techniques to accurately predict the aerodynamic interactions between heterogeneous multirotors: 1) Spectral normalization for stability and generalization guarantees of unseen data and 2) heterogeneous deep sets for supporting any number of heterogeneous neighbors in a permutation-invariant manner without reducing expressiveness. The learned residual dynamics benefit both the proposed interaction-aware multirobot motion planning and the nonlinear tracking control design because the learned interaction forces reduce the modelling errors. Experimental results demonstrate that Neural-Swarm2 is able to generalize to larger swarms beyond training cases and significantly outperforms a baseline nonlinear tracking controller with up to three times reduction in worst-case tracking errors.", "date": "2022-04", "date_type": "published", "publication": "IEEE Transactions on Robotics", "volume": "38", "number": "2", "publisher": "IEEE", "pagerange": "1063-1079", "id_number": "CaltechAUTHORS:20210120-165259145", "issn": "1552-3098", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20210120-165259145", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Center for Autonomous Systems and Technologies" }, { "agency": "Raytheon Company" }, { "agency": "JPL" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" }, { "id": "GALCIT" } ] }, "doi": "10.1109/TRO.2021.3098436", "primary_object": { "basename": "2012.05457.pdf", "url": "https://authors.library.caltech.edu/records/psxc9-vps89/files/2012.05457.pdf" }, "pub_year": "2022", "author_list": "Shi, Guanya; H\u00f6nig, Wolfgang; et al." }, { "id": "https://authors.library.caltech.edu/records/wmv09-93s03", "eprint_id": 113939, "eprint_status": "archive", "datestamp": "2023-08-22 14:51:16", "lastmod": "2025-11-20 22:47:14", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Daftry-Shreyansh", "name": { "family": "Daftry", "given": "Shreyansh" }, "orcid": "0000-0002-0157-5944" }, { "id": "Abcouwer-Neil", "name": { "family": "Abcouwer", "given": "Neil" }, "orcid": "0000-0002-1346-6294" }, { "id": "del-Sesto-Tyler", "name": { "family": "del Sesto", "given": "Tyler" } }, { "id": "Venkatraman-Siddarth", "name": { "family": "Venkatraman", "given": "Siddarth" }, "orcid": "0000-0002-3607-2781" }, { "id": "Song-Jialin", "name": { "family": "Song", "given": "Jialin" } }, { "id": "Igel-Lucas", "name": { "family": "Igel", "given": "Lucas" } }, { "id": "Byon-Amos", "name": { "family": "Byon", "given": "Amos" } }, { "id": "Rosolia-Ugo", "name": { "family": "Rosolia", "given": "Ugo" }, "orcid": "0000-0002-1682-0551" }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" }, { "id": "Ono-Masahiro", "name": { "family": "Ono", "given": "Masahiro" }, "orcid": "0000-0002-9247-1306" } ] }, "title": "MLNav: Learning to Safely Navigate on Martian Terrains", "ispublished": "pub", "full_text_status": "public", "keywords": "Integrated planning and learning, motion andpath planning, space robotics and automation; Artificial Intelligence; Control and Optimization; Computer Science Applications; Computer Vision and Pattern Recognition; Mechanical Engineering; Human-Computer Interaction; Biomedical Engineering; Control and Systems Engineering", "note": "\u00a9 2022 IEEE. \n\nManuscript received September 9, 2021; accepted February 13, 2022. Date of publication March 7, 2022; date of current version March 16, 2022. This letter was recommended for publication by Associate Editor Chris Paxton and Editor Stephen J. Guy upon evaluation of the reviewers' comments. \n\nThis work was supported by the JPL Research and Technology Development (R&TD) program. This work was supported in-part by Raytheon. \n\nThis work was supported by the Jet Propulsion Laboratory, California Institute of Technology, and California Institute of Technology under a contract with the National Aeronautics and Space Administration. The authors would like to thank Olivier Toupet, Mitch Ingham and Ravi Lanka for valuable discussions and problem formulation.\n\nAccepted Version - 2203.04563.pdf
", "abstract": "We present MLNav, a learning-enhanced path planning framework for safety-critical and resource-limited systems operating in complex environments, such as rovers navigating on Mars. MLNav makes judicious use of machine learning to enhance the efficiency of path planning while fully respecting safety constraints. In particular, the dominant computational cost in such safety-critical settings is running a model-based safety checker on the proposed paths. Our learned search heuristic can simultaneously predict the feasibility for all path options in a single run, and the model-based safety checker is only invoked on the top-scoring paths. We validate in high-fidelity simulations using both real Martian terrain data collected by the Perseverance rover, as well as a suite of challenging synthetic terrains. Our experiments show that: (i) compared to the baseline ENav path planner on board the Perserverance rover, MLNav can provide a significant improvement in multiple key metrics, such as a 10x reduction in collision checks when navigating real Martian terrains, despite being trained with synthetic terrains; and (ii) MLNav can successfully navigate highly challenging terrains where the baseline ENav fails to find a feasible path before timing out.", "date": "2022-04", "date_type": "published", "publication": "IEEE Robotics and Automation Letters", "volume": "7", "number": "2", "publisher": "IEEE", "pagerange": "5461-5468", "id_number": "CaltechAUTHORS:20220317-376217000", "issn": "2377-3766", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220317-376217000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "JPL Research and Technology Development Fund" }, { "agency": "Raytheon Company" }, { "agency": "NASA/JPL/Caltech" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/lra.2022.3156654", "primary_object": { "basename": "2203.04563.pdf", "url": "https://authors.library.caltech.edu/records/wmv09-93s03/files/2203.04563.pdf" }, "pub_year": "2022", "author_list": "Daftry, Shreyansh; Abcouwer, Neil; et al." }, { "id": "https://authors.library.caltech.edu/records/h67h9-yw208", "eprint_id": 113945, "eprint_status": "archive", "datestamp": "2023-10-09 20:59:47", "lastmod": "2025-11-21 00:25:36", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Veismann-Marcel", "name": { "family": "Veismann", "given": "Marcel" }, "orcid": "0000-0001-8106-6738" }, { "id": "Yos-Daniel", "name": { "family": "Yos", "given": "Daniel" } }, { "id": "Gharib-M", "name": { "family": "Gharib", "given": "Morteza" }, "orcid": "0000-0003-0754-4193" } ] }, "title": "Parametric study of small-scale rotors in axial descent", "ispublished": "pub", "full_text_status": "public", "keywords": "Condensed Matter Physics; Fluid Flow and Transfer Processes; Mechanics of Materials; Computational Mechanics; Mechanical Engineering", "note": "\u00a9 2022 Author(s). Published under an exclusive license by AIP Publishing. \n\nSubmitted: 29 December 2021 . Accepted: 25 February 2022 . Published Online: 17 March 2022. \n\nThis research was funded by the Center of Autonomous Systems and Technology (CAST) and the Summer Undergraduate Research Fellowship (SURF) at the California Institute of Technology. \n\nDATA AVAILABILITY. The data that support the findings of this study are available from the corresponding author upon reasonable request. \n\nThe authors have no conflicts to disclose.\n\nPublished - 035124_1_online.pdf
Supplemental Material - supplementary_material.pdf
", "abstract": "Despite extensive research in multirotor aerodynamics in the recent past, axial descent, specifically the vortex ring state, still poses great challenges for multirotor configurations as this flight stage is typically accompanied by severe losses in rotor thrust and strong thrust fluctuations. This paper presents a parametric study to investigate the influence of relevant geometric parameters of a small-scale rotor blade on the rotor performance in axial descent. Design variables subject to variation were the collective pitch, chord length, taper ratio, number of blades, as well as the tip geometry. Custom rotors for each parameter modification were manufactured and experimentally evaluated in wind tunnel tests with mean thrust recordings and measurements of the thrust fluctuations serving as performance metrics. Results indicated that rotor blades with larger aspect ratio and higher blade loading coefficient are less affected by the adverse aerodynamics in the vortex ring state, experiencing lower thrust losses and vibrational loads. Particle image velocimetry flow visualization confirmed that the aerodynamic losses in the vortex ring state can be attributed to blade vortex interactions. Comparison of the rotor flow structure in hover of all investigated rotor designs suggested that improvements in the descent performance of a rotor stem from a combination of reduced tip vortex strength and increased axial tip vortex convection rate. Using the experimental findings of this study, a predictive model for approximating the maximum extent of mean thrust losses in axial descent for a given blade geometry and hover thrust coefficient could be established.", "date": "2022-03", "date_type": "published", "publication": "Physics of Fluids", "volume": "34", "number": "3", "publisher": "American Institute of Physics", "pagerange": "Art. No. 035124", "id_number": "CaltechAUTHORS:20220317-376350000", "issn": "1070-6631", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220317-376350000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Center for Autonomous Systems and Technologies" }, { "agency": "Caltech Summer Undergraduate Research Fellowship (SURF)" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1063/5.0083761", "primary_object": { "basename": "supplementary_material.pdf", "url": "https://authors.library.caltech.edu/records/h67h9-yw208/files/supplementary_material.pdf" }, "related_objects": [ { "basename": "035124_1_online.pdf", "url": "https://authors.library.caltech.edu/records/h67h9-yw208/files/035124_1_online.pdf" } ], "pub_year": "2022", "author_list": "Veismann, Marcel; Yos, Daniel; et al." }, { "id": "https://authors.library.caltech.edu/records/rmv3t-7zv14", "eprint_id": 113029, "eprint_status": "archive", "datestamp": "2023-08-22 10:36:06", "lastmod": "2025-11-20 23:39:44", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Rivi\u00e8re-Benjamin", "name": { "family": "Rivi\u00e8re", "given": "Benjamin" }, "orcid": "0000-0003-4189-4090" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "H-TD\u00b2: Hybrid Temporal Difference Learning for Adaptive Urban Taxi Dispatch", "ispublished": "pub", "full_text_status": "public", "keywords": "Real-time taxi dispatch, adaptive systems, multi-agent systems, distributed decision-making, autonomous\nvehicles; Computer Science Applications; Mechanical Engineering; Automotive Engineering", "note": "\u00a9 2021 IEEE. \n\nManuscript received September 4, 2019; revised April 20, 2020 and April 22, 2021; accepted June 15, 2021. \n\nThis work was supported in part by Raytheon Company and in part by the Jet Propulsion Laboratory. \n\nThe authors would like to thank Salar Rahili, who proposed the use of binary log-linear learning in an early version of this work.\n\nAccepted Version - H-TD_Hybrid_Temporal_Difference_Learning_for_Adaptive_Urban_Taxi_Dispatch.pdf
Submitted - 2105.02138.pdf
", "abstract": "We present H-TD\u00b2: Hybrid Temporal Difference Learning for Taxi Dispatch, a model-free, adaptive decision-making algorithm to coordinate a large fleet of automated taxis in a dynamic urban environment to minimize expected customer waiting times. Our scalable algorithm exploits the natural transportation network company topology by switching between two behaviors: distributed temporal-difference learning computed locally at each taxi and infrequent centralized Bellman updates computed at the dispatch center. We derive a regret bound and design the trigger condition between the two behaviors to explicitly control the trade-off between computational complexity and the individual taxi policy's bounded sub-optimality; this advances the state of the art by enabling distributed operation with bounded-suboptimality. Additionally, unlike recent reinforcement learning dispatch methods, this policy estimation is adaptive and robust to out-of-training domain events. This result is enabled by a two-step modelling approach: the policy is learned on an agent-agnostic, cell-based Markov Decision Process and individual taxis are coordinated using the learned policy in a distributed game-theoretic task assignment. We validate our algorithm against a receding horizon control baseline in a Gridworld environment with a simulated customer dataset, where the proposed solution decreases average customer waiting time by 50% over a wide range of parameters. We also validate in a Chicago city environment with real customer requests from the Chicago taxi public dataset where the proposed solution decreases average customer waiting time by 26% over irregular customer distributions during a 2016 Major League Baseball World Series game.", "date": "2022-01-20", "date_type": "published", "publication": "IEEE Transactions on Intelligent Transportation Systems", "publisher": "IEEE", "id_number": "CaltechAUTHORS:20220120-890613000", "issn": "1524-9050", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220120-890613000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Raytheon Company" }, { "agency": "JPL" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/tits.2021.3097297", "primary_object": { "basename": "2105.02138.pdf", "url": "https://authors.library.caltech.edu/records/rmv3t-7zv14/files/2105.02138.pdf" }, "related_objects": [ { "basename": "H-TD_Hybrid_Temporal_Difference_Learning_for_Adaptive_Urban_Taxi_Dispatch.pdf", "url": "https://authors.library.caltech.edu/records/rmv3t-7zv14/files/H-TD_Hybrid_Temporal_Difference_Learning_for_Adaptive_Urban_Taxi_Dispatch.pdf" } ], "pub_year": "2022", "author_list": "Rivi\u00e8re, Benjamin and Chung, Soon-Jo" }, { "id": "https://authors.library.caltech.edu/records/vbg0f-dts09", "eprint_id": 113279, "eprint_status": "archive", "datestamp": "2023-10-09 20:55:41", "lastmod": "2023-10-24 16:29:00", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Wei-Nathaniel-J", "name": { "family": "Wei", "given": "Nathaniel J." }, "orcid": "0000-0001-5846-6485" }, { "id": "Dabiri-J-O", "name": { "family": "Dabiri", "given": "John O." }, "orcid": "0000-0002-6722-9008" } ] }, "title": "Phase-averaged dynamics of a periodically surging wind turbine", "ispublished": "pub", "full_text_status": "public", "keywords": "Renewable Energy, Sustainability and the Environment", "note": "\u00a9 2022 Author(s). Published under an exclusive license by AIP Publishing. \n\nSubmitted: 20 October 2021. Accepted: 17 December 2021. Accepted Manuscript Online: 17 December 2021. Published Online: 02 February 2022. \n\nThe authors gratefully acknowledge the assistance of several people, without whom the construction and operation of the experimental apparatus would not have been possible: J. Benson, who graciously provided machine-shop access during the pandemic; G. Juarez and M. Vega, who installed the power systems for the linear actuator; M. Miller, K. Bankord, and J. Kissing for technical consultation regarding the turbine power-control system and linear actuator; A. Kiani and E. Tang for machining key components of the apparatus; N. Esparza-Duran and R. Nemovi, who oversaw operations at CAST; M. Veismann and P. Renn for wind-tunnel support; and J. Cardona, E. Tang, P. Gunnarson, M. Fu, and R. Goldshmid for providing assistance and safety supervision for the experiments. \n\nThis work was funded by the National Science Foundation (Grant No. CBET-2038071) and the Caltech Center for Autonomous Systems and Technologies. N. Wei was supported by a Stanford Graduate Fellowship and a National Science Foundation Graduate Research Fellowship. \n\nDATA AVAILABILITY. The data that support the findings of this study are available from the corresponding author upon reasonable request. \n\nThe authors have no conflicts to disclose.\n\nPublished - 013305_1_online.pdf
Submitted - 2110-10312.pdf
", "abstract": "The unsteady power generation of a wind turbine translating in the streamwise direction is relevant to floating offshore wind turbines, kite-mounted airborne wind turbines, and other non-traditional wind-energy systems. To study this problem experimentally, measurements of torque, rotor speed, and power were acquired for a horizontal-axis wind turbine actuated in periodic surge motions in a fan-array wind tunnel at the Caltech Center for Autonomous Systems and Technologies (CAST). Experiments were conducted at a diameter-based Reynolds number of Re_D = 6.1 \u00d7 10\u2075 and at tip-speed ratios between 5.2 and 8.8. Sinusoidal and trapezoidal surge-velocity waveforms with maximum surge velocities up to 23% of the free-stream velocity were tested. A model in the form of a linear ordinary differential equation (first-order in time) was derived to capture the time-resolved dynamics of the surging turbine. Its coefficients were obtained using torque measurements from a stationary turbine, without the need for unsteady calibrations. Its predictions compared favorably with the measured amplitude- and phase-response data. Furthermore, increases in the period-averaged power of up to 6.4% above the steady reference case were observed in the experiments at high tip-speed ratios and surge velocities, potentially due to unsteady or nonlinear aerodynamic effects. Conversely, decreases in mean power with increased surge velocity at low tip-speed ratios were likely a result of the onset of stall on the turbine blades. These results inform the development of strategies to optimize and control the unsteady power generation of periodically surging wind turbines, and motivate further investigations into the unsteady aerodynamics of wind-energy systems.", "date": "2022-01", "date_type": "published", "publication": "Journal of Renewable and Sustainable Energy", "volume": "14", "number": "1", "publisher": "American Institute of Physics", "pagerange": "Art. No. 013305", "id_number": "CaltechAUTHORS:20220204-350702000", "issn": "1941-7012", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220204-350702000", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "NSF", "grant_number": "CBET-2038071" }, { "agency": "Center for Autonomous Systems and Technologies" }, { "agency": "Stanford University" }, { "agency": "NSF Graduate Research Fellowship" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1063/5.0076029", "primary_object": { "basename": "2110-10312.pdf", "url": "https://authors.library.caltech.edu/records/vbg0f-dts09/files/2110-10312.pdf" }, "related_objects": [ { "basename": "013305_1_online.pdf", "url": "https://authors.library.caltech.edu/records/vbg0f-dts09/files/013305_1_online.pdf" } ], "pub_year": "2022", "author_list": "Wei, Nathaniel J. and Dabiri, John O." }, { "id": "https://authors.library.caltech.edu/records/yp1dr-27k89", "eprint_id": 109030, "eprint_status": "archive", "datestamp": "2023-08-20 06:22:38", "lastmod": "2024-03-27 23:54:58", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Taylor-Andrew-J", "name": { "family": "Taylor", "given": "Andrew J." }, "orcid": "0000-0002-5990-590X" }, { "id": "Dorobantu-Victor-D", "name": { "family": "Dorobantu", "given": "Victor D." }, "orcid": "0000-0002-2797-7802" }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" }, { "id": "Tabuada-Paulo", "name": { "family": "Tabuada", "given": "Paulo" }, "orcid": "0000-0002-3417-0951" }, { "id": "Ames-A-D", "name": { "family": "Ames", "given": "Aaron D." }, "orcid": "0000-0003-0848-3177" } ] }, "title": "Sampled-Data Stabilization with Control Lyapunov Functions via Quadratically Constrained Quadratic Programs", "ispublished": "pub", "full_text_status": "public", "keywords": "Sampled-data control, Lyapunov methods, discrete event systems", "note": "\u00a9 2021 IEEE. \n\nManuscript received March 4, 2021; revised May 10, 2021; accepted May 16, 2021. Date of publication June 2, 2021; date of current version June 29, 2021. \n\nThe work of Andrew J. Taylor and Aaron D. Ames was supported by NSF under Award 1932091. The work of Victor D. Dorobantu and Yisong Yue was supported in part by DARPA, in part by Beyond Limits, and in part by a Kortschak Fellowship. The work of Paulo Tabuada was supported in part by NSF under Award 1705135. \n\nRecommended by Senior Editor L. Menini.\n\nSubmitted - 2103.03937.pdf
", "abstract": "Controller design for nonlinear systems with Control Lyapunov Function (CLF) based quadratic programs has recently been successfully applied to a diverse set of difficult control tasks. These existing formulations do not address the gap between design with continuous time models and the discrete time sampled implementation of the resulting controllers, often leading to poor performance on hardware platforms. We propose an approach to close this gap by synthesizing sampled-data counterparts to these CLF-based controllers, specified as quadratically constrained quadratic programs (QCQPs). Assuming feedback linearizability and stable zero-dynamics of a system's continuous time model, we derive practical stability guarantees for the resulting sampled-data system. We demonstrate improved performance of the proposed approach over continuous time counterparts in simulation.", "date": "2022", "date_type": "published", "publication": "IEEE Control Systems Letters", "volume": "6", "publisher": "IEEE", "pagerange": "680-685", "id_number": "CaltechAUTHORS:20210510-100142166", "issn": "2475-1456", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20210510-100142166", "funders": { "items": [ { "agency": "NSF", "grant_number": "CNS-1932091" }, { "agency": "Defense Advanced Research Projects Agency (DARPA)" }, { "agency": "Beyond Limits" }, { "agency": "Kortschak Scholars Program" }, { "agency": "NSF", "grant_number": "CNS-1705135" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/LCSYS.2021.3085172", "primary_object": { "basename": "2103.03937.pdf", "url": "https://authors.library.caltech.edu/records/yp1dr-27k89/files/2103.03937.pdf" }, "pub_year": "2022", "author_list": "Taylor, Andrew J.; Dorobantu, Victor D.; et al." }, { "id": "https://authors.library.caltech.edu/records/2wmvb-ah673", "eprint_id": 107467, "eprint_status": "archive", "datestamp": "2023-08-20 06:04:49", "lastmod": "2026-03-30 07:23:07", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Folkestad-Carl", "name": { "family": "Folkestad", "given": "Carl" }, "orcid": "0000-0002-3436-8247" }, { "id": "Chen-Yuxiao", "name": { "family": "Chen", "given": "Yuxiao" }, "orcid": "0000-0001-5276-7156" }, { "id": "Ames-A-D", "name": { "family": "Ames", "given": "Aaron D." }, "orcid": "0000-0003-0848-3177" }, { "id": "Burdick-J-W", "name": { "family": "Burdick", "given": "Joel W." }, "orcid": "0000-0002-3091-540X" } ] }, "title": "Data-Driven Safety-Critical Control: Synthesizing Control Barrier Functions With Koopman Operators", "ispublished": "pub", "full_text_status": "restricted", "keywords": "Robotics, computational methods, supervisory\ncontrol", "note": "\u00a9 2020 IEEE. \n\nManuscript received September 14, 2020; revised November\n21, 2020; accepted December 11, 2020. Date of publication\nDecember 21, 2020; date of current version March 22, 2021.\n\nThis work was supported in part by Raytheon Technologies. The work of Carl Folkestad was supported by the Aker Scholarship Foundation. Recommended by Senior Editor F. Dabben.", "abstract": "Control barrier functions (CBFs) are a powerful tool to guarantee safety of autonomous systems, yet they rely on the computation of control invariant sets, which is notoriously difficult. A backup strategy employs an implicit control invariant set computed by forward integrating the system dynamics. However, this integration is prohibitively expensive for high dimensional systems, and inaccurate in the presence of unmodelled dynamics. We propose to learn discrete-time Koopman operators of the closed-loop dynamics under a backup strategy. This approach replaces forward integration by a simple matrix multiplication, which can mostly be computed offline. We also derive an error bound on the unmodeled dynamics in order to robustify the CBF controller. Our approach extends to multi-agent systems, and we demonstrate the method on collision avoidance for wheeled robots and quadrotors.", "date": "2021-12", "date_type": "published", "publication": "IEEE Control Systems Letters", "volume": "5", "number": "6", "publisher": "IEEE", "pagerange": "2012-2017", "id_number": "CaltechAUTHORS:20210113-163505361", "issn": "2475-1456", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20210113-163505361", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Raytheon Company" }, { "agency": "Aker Scholarship Foundation" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/lcsys.2020.3046159", "pub_year": "2021", "author_list": "Folkestad, Carl; Chen, Yuxiao; et al." }, { "id": "https://authors.library.caltech.edu/records/ndcb6-z5n54", "eprint_id": 106580, "eprint_status": "archive", "datestamp": "2023-08-22 11:59:47", "lastmod": "2026-03-28 22:27:50", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Yoo-Gene-Ryan", "name": { "family": "Yoo", "given": "Gene Ryan" }, "orcid": "0000-0002-5319-5599" }, { "id": "Owhadi-H", "name": { "family": "Owhadi", "given": "Houman" }, "orcid": "0000-0002-5677-1600" } ] }, "title": "Deep regularization and direct training of the inner layers of Neural Networks with Kernel Flows", "ispublished": "pub", "full_text_status": "public", "keywords": "Kernel Flows; Gaussian process regression; Artificial neural networks; Machine learning; Image classification; Inner layer training", "note": "\u00a9 2021 Published by Elsevier B.V. \n\nReceived 31 December 2020, Revised 17 March 2021, Accepted 13 April 2021, Available online 18 July 2021. \n\nThe authors gratefully acknowledge support by the Air Force Office of Scientific Research, USA under award number FA9550-18-1-0271 (Games for Computation and Learning), Beyond Limits, USA (Learning Optimal Models), and NASA/JPL, USA (Earth 2050). We also thank an anonymous referee for comments and suggestions. \n\nCRediT authorship contribution statement: Gene Ryan Yoo: Conceptualization, Methodology, Numerical experiments, Writing \u2013 original draft. Houman Owhadi: Conceptualization, Writing \u2013 review & editing. \n\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\n\nSubmitted - 2002.08335.pdf
", "abstract": "We introduce a new regularization method for Artificial Neural Networks (ANNs) based on the Kernel Flow (KF) algorithm. The algorithm was introduced in Owhadi and Yoo (2019) as a method for kernel selection in regression/kriging based on the minimization of the loss of accuracy incurred by halving the number of interpolation points in random batches of the dataset. Writing f_\u03b8(x) = (f^((n))_(\u03b8n)\u2218f^((n\u22121))_(\u03b8n\u22121)\u2218\u22ef\u2218f^((1))_(\u03b8\u2081))(x) for the functional representation of compositional structure of the ANN (where \u03b8_i are the weights and biases of the layer i), the inner layers outputs h^((i))(x) = (f^((i))_(\u03b8i)\u2218f^((i\u22121))_(\u03b8i\u22121)\u2218\u22ef\u2218f^((1))_(\u03b81))(x) define a hierarchy of feature maps and a hierarchy of kernels k^((i))(x,x\u2032) = exp(\u2212\u03b3_i\u2225h^((i))(x)\u2212h^((i))(x\u2032)\u2225\u00b2\u2082). When combined with a batch of the dataset, these kernels produce KF losses e(i)\u2082 (defined as the L\u00b2 regression error incurred by using a random half of the batch to predict the other half) depending on the parameters of the inner layers \u03b8\u2081,\u2026,\u03b8_i (and \u03b3_i). The proposed method simply consists of aggregating (as a weighted sum) a subset of these KF losses with a classical output loss (e.g., cross-entropy). We test the proposed method on Convolutional Neural Networks (CNNs) and Wide Residual Networks (WRNs) without alteration of their structure nor their output classifier and report reduced test errors, decreased generalization gaps, and increased robustness to distribution shift without a significant increase in computational complexity relative to standard CNN and WRN training (with Drop Out and Batch Normalization). We suspect that these results might be explained by the fact that while conventional training only employs a linear functional (a generalized moment) of the empirical distribution defined by the dataset and can be prone to trapping in the Neural Tangent Kernel regime (under over-parameterizations), the proposed loss function (defined as a nonlinear functional of the empirical distribution) effectively trains the underlying kernel defined by the CNN beyond regressing the data with that kernel.", "date": "2021-11-15", "date_type": "published", "publication": "Physica D: Nonlinear Phenomena", "volume": "426", "publisher": "Elsevier", "pagerange": "Art. No. 132952", "id_number": "CaltechAUTHORS:20201110-075343797", "issn": "0167-2789", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20201110-075343797", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Air Force Office of Scientific Research (AFOSR)", "grant_number": "FA9550-18-1-0271" }, { "agency": "NASA/JPL", "grant_number": "Earth 2050" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1016/j.physd.2021.132952", "primary_object": { "basename": "2002.08335.pdf", "url": "https://authors.library.caltech.edu/records/ndcb6-z5n54/files/2002.08335.pdf" }, "pub_year": "2021", "author_list": "Yoo, Gene Ryan and Owhadi, Houman" }, { "id": "https://authors.library.caltech.edu/records/9b7jr-g4j06", "eprint_id": 107468, "eprint_status": "archive", "datestamp": "2023-08-20 05:45:45", "lastmod": "2026-03-30 16:05:34", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Tsukamoto-Hiroyasu", "name": { "family": "Tsukamoto", "given": "Hiroyasu" } }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" }, { "id": "Slotine-Jean-Jacques-E", "name": { "family": "Slotine", "given": "Jean-Jacques E." }, "orcid": "0000-0002-7161-7812" } ] }, "title": "Neural Stochastic Contraction Metrics for Learning-based Control and Estimation", "ispublished": "pub", "full_text_status": "public", "keywords": "Machine learning, stochastic optimal control, observers for nonlinear systems", "note": "\u00a9 2020 IEEE. \n\nManuscript received September 14, 2020; revised November\n18, 2020; accepted December 7, 2020. Date of publication\nDecember 22, 2020; date of current version January 13, 2021. \n\nThis work was supported in part by the Raytheon Company. \n\nThis work was benefited from discussions with Nicholas Boffi and Quang-Cuong Pham. Code: https://github.com/astrohiro/nscm.\n\nAccepted Version - 2011.03168.pdf
", "abstract": "We present Neural Stochastic Contraction Metrics (NSCM), a new design framework for provably-stable learning-based control and estimation for a class of stochastic nonlinear systems. It uses a spectrally-normalized deep neural network to construct a contraction metric and its differential Lyapunov function, sampled via simplified convex optimization in the stochastic setting. Spectral normalization constrains the state-derivatives of the metric to be Lipschitz continuous, thereby ensuring exponential boundedness of the mean squared distance of system trajectories under stochastic disturbances. The trained NSCM model allows autonomous systems to approximate optimal stable control and estimation policies in real-time, and outperforms existing nonlinear control and estimation techniques including the state-dependent Riccati equation, iterative LQR, EKF, and the deterministic NCM, as shown in simulation results.", "date": "2021-11", "date_type": "published", "publication": "IEEE Control Systems Letters", "volume": "5", "number": "5", "publisher": "IEEE", "pagerange": "1825-1830", "id_number": "CaltechAUTHORS:20210113-163505450", "issn": "2475-1456", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20210113-163505450", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Raytheon Company" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/lcsys.2020.3046529", "primary_object": { "basename": "2011.03168.pdf", "url": "https://authors.library.caltech.edu/records/9b7jr-g4j06/files/2011.03168.pdf" }, "pub_year": "2021", "author_list": "Tsukamoto, Hiroyasu; Chung, Soon-Jo; et al." }, { "id": "https://authors.library.caltech.edu/records/nse0n-x2w89", "eprint_id": 106456, "eprint_status": "archive", "datestamp": "2023-08-20 05:45:23", "lastmod": "2026-03-29 14:42:23", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Ahmadi-Mohamadreza", "name": { "family": "Ahmadi", "given": "Mohamadreza" }, "orcid": "0000-0003-1447-3012" }, { "id": "Jansen-Nils", "name": { "family": "Jansen", "given": "Nils" }, "orcid": "0000-0003-1318-8973" }, { "id": "Wu-Bo", "name": { "family": "Wu", "given": "Bo" }, "orcid": "0000-0002-7199-6525" }, { "id": "Topcu-Ufuk", "name": { "family": "Topcu", "given": "Ufuk" }, "orcid": "0000-0003-0819-9985" } ] }, "title": "Control Theory Meets POMDPs: A Hybrid Systems Approach", "ispublished": "pub", "full_text_status": "public", "keywords": "Artificial intelligence, autonomous systems, control theory, Lyapunov methods, Markov processes", "note": "\u00a9 2020 IEEE. \n\nManuscript received April 15, 2020; accepted October 28, 2020. Date of publication November 4, 2020; date of current version November 4, 2021. \n\nThis work was supported by Grant AFOSR FA9550-19-1-0005, Grant AFRL FA9550-19-1-0169, Grant DARPA D19AP00004, Grant NSF 1646522, Grant NWO OCENW.KLEIN.187, and Grant NSF 1652113. Recommended by Associate Editor L. Palopoli. \n\nM. Ahmadi appreciates the stimulating discussions with Dr.Y.\nChen at the University of Chicago, Prof. Y. Yue at Caltech, and Prof. R. M. Murray at Caltech.\n\nAccepted Version - 09248580.pdf
Submitted - 1905.08095.pdf
", "abstract": "Partially observable Markov decision processes (POMDPs) provide a modeling framework for a variety of sequential decision making under uncertainty scenarios in artificial intelligence (AI). Since the states are not directly observable in a POMDP, decision making has to be performed based on the output of a Bayesian filter (continuous beliefs); hence, making POMDPs intractable to solve and analyze. To overcome the complexity challenge of POMDPs, we apply techniques from the control theory. Our contributions are fourfold. 1) We begin by casting the problem of analyzing a POMDP into analyzing the behavior of a discrete-time switched system. 2) Then, in order to estimate the reachable belief space of a POMDP, i.e., the set of all possible evolutions given an initial belief distribution over the states and a set of actions and observations, we find overapproximations in terms of sublevel sets of Lyapunov-like functions. 3) Furthermore, in order to verify safety and performance requirements of a given POMDP, we formulate a barrier certificate theorem, wherein we show that if there exists a barrier certificate satisfying a set of inequalities along the solutions to the belief update equation of the POMDP, the safety and performance properties are guaranteed to hold. In both cases 2) and 3), the calculations can be decomposed and solved in parallel. 4) Finally, we show that the conditions we formulate can be computationally implemented as a set of sum-of-squares programs. We illustrate the applicability of our method by addressing two problems in active ad scheduling and machine teaching.", "date": "2021-11", "date_type": "published", "publication": "IEEE Transactions on Automatic Control", "volume": "66", "number": "11", "publisher": "IEEE", "pagerange": "5191-5204", "id_number": "CaltechAUTHORS:20201105-145616123", "issn": "0018-9286", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20201105-145616123", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Air Force Office of Scientific Research (AFOSR)", "grant_number": "FA9550-19-1-0005" }, { "agency": "Air Force Research Laboratory (AFRL)", "grant_number": "FA9550-19-1-0169" }, { "agency": "Defense Advanced Research Projects Agency (DARPA)", "grant_number": "D19AP00004" }, { "agency": "NSF", "grant_number": "CNS-1646522" }, { "agency": "Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO)", "grant_number": "OCENW.KLEIN.187" }, { "agency": "NSF", "grant_number": "CNS-1652113" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/tac.2020.3035755", "primary_object": { "basename": "09248580.pdf", "url": "https://authors.library.caltech.edu/records/nse0n-x2w89/files/09248580.pdf" }, "related_objects": [ { "basename": "1905.08095.pdf", "url": "https://authors.library.caltech.edu/records/nse0n-x2w89/files/1905.08095.pdf" } ], "pub_year": "2021", "author_list": "Ahmadi, Mohamadreza; Jansen, Nils; et al." }, { "id": "https://authors.library.caltech.edu/records/a3y6x-12v73", "eprint_id": 111265, "eprint_status": "archive", "datestamp": "2023-08-20 05:30:10", "lastmod": "2026-03-30 15:30:05", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Kim-Kyunam", "name": { "family": "Kim", "given": "Kyunam" }, "orcid": "0000-0002-7803-1582" }, { "id": "Spieler-Patrick", "name": { "family": "Spieler", "given": "Patrick" } }, { "id": "Lupu-Elena-Sorina", "name": { "family": "Lupu", "given": "Elena Sorina" }, "orcid": "0000-0002-3968-2630" }, { "id": "Ramezani-Alireza", "name": { "family": "Ramezani", "given": "Alireza" }, "orcid": "0000-0002-3391-5288" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "A bipedal walking robot that can fly, slackline, and skateboard", "ispublished": "pub", "full_text_status": "public", "note": "\u00a9 2021 The Authors, some rights reserved; exclusive licensee\nAmerican Association for the Advancement of Science. No claim to original U.S. Government Works. \n\nSubmitted 16 December 2020; Accepted 8 September 2021; Published 6 October 2021. \n\nWe thank Y. Veys, S. van Nieuwstadt, and B. Cruz for contributing to an early phase of design and control work, as well as N. Esparza-Duran for heel manufacturing. \n\nThis work was in part funded by the Caltech Gary Clinard Innovation Fund. We thank M. Gharib and the Center for Autonomous Systems and Technologies for the funding support. \n\nAuthor contributions: A.R. and S.-J.C. conceived and envisioned a first prototype of LEO, which was developed by A.R. with critical feedback and input from S.-J.C. S.-J.C. directed the research activities that enabled demonstration of the new LEO robot concept reported in the article. K.K. and P.S. designed the robot, and P.S. selected the main components, built the robot hardware, and implemented the software. K.K., P.S., and E.-S.L. contributed to modeling, controller design and stability proof, and performance analysis of LEO with critical feedback and input from S.-J.C. K.K., P.S., E.-S.L., and S.-J.C. designed experiment plans. P.S. and E.-S.L. performed experiments, and K.K., P.S., and E.-S.L. evaluated experimental results. K.K., P.S., E.-S.L., and S.-J.C. prepared and edited the manuscript, and all authors reviewed the manuscript. \n\nCompeting interests: California Institute of Technology filed a U.S. nonprovisional patent application on this work on 23 December 2020. \n\nData and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.\n\nSupplemental Material - scirobotics.abf8136_movies_s1_to_s8.zip
Supplemental Material - scirobotics.abf8136_sm.pdf
", "abstract": "Numerous mobile robots in various forms specialize in either ground or aerial locomotion, whereas very few robots can perform complex locomotion tasks beyond simple walking and flying. We present the design and control of a multimodal locomotion robotic platform called LEONARDO, which bridges the gap between two different locomotion regimes of flying and walking using synchronized control of distributed electric thrusters and a pair of multijoint legs. By combining two distinct locomotion mechanisms, LEONARDO achieves complex maneuvers that require delicate balancing, such as walking on a slackline and skateboarding, which are challenging for existing bipedal robots. LEONARDO also demonstrates agile walking motions, interlaced with flying maneuvers to overcome obstacles using synchronized control of propellers and leg joints. The mechanical design and synchronized control strategy achieve a unique multimodal locomotion capability that could potentially enable robotic missions and operations that would be difficult for single-modal locomotion robots.", "date": "2021-10-06", "date_type": "published", "publication": "Science Robotics", "volume": "6", "number": "59", "publisher": "American Association for the Advancement of Science", "pagerange": "Art. No. eabf8136", "id_number": "CaltechAUTHORS:20211007-153559085", "issn": "2470-9476", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20211007-153559085", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Caltech Gary Clinard Innovation Fund" }, { "agency": "Center for Autonomous Systems and Technologies" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" }, { "id": "GALCIT" } ] }, "doi": "10.1126/scirobotics.abf8136", "primary_object": { "basename": "scirobotics.abf8136_movies_s1_to_s8.zip", "url": "https://authors.library.caltech.edu/records/a3y6x-12v73/files/scirobotics.abf8136_movies_s1_to_s8.zip" }, "related_objects": [ { "basename": "scirobotics.abf8136_sm.pdf", "url": "https://authors.library.caltech.edu/records/a3y6x-12v73/files/scirobotics.abf8136_sm.pdf" } ], "pub_year": "2021", "author_list": "Kim, Kyunam; Spieler, Patrick; et al." }, { "id": "https://authors.library.caltech.edu/records/38h8h-egc30", "eprint_id": 99546, "eprint_status": "archive", "datestamp": "2023-08-20 05:15:40", "lastmod": "2026-03-29 21:49:42", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Tsukamoto-Hiroyasu", "name": { "family": "Tsukamoto", "given": "Hiroyasu" }, "orcid": "0000-0002-6337-2667" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Robust Controller Design for Stochastic Nonlinear Systems via Convex Optimization", "ispublished": "pub", "full_text_status": "public", "keywords": "LMIs, nonlinear systems, optimization algorithms, robust control, stochastic optimal control", "note": "\u00a9 2020 IEEE. \n\nManuscript received June 4, 2020; accepted October 28, 2020. Date of publication November 16, 2020; date of current version September 27, 2021. \n\nThis work was supported in part by the Jet Propulsion Laboratory, California Institute of Technology and in part by the Raytheon Company. Recommended by Associate Editor U. V. Shanbhag.\n\nAccepted Version - 09261103.pdf
Submitted - 2006.04359.pdf
", "abstract": "This article presents ConVex optimization-based Stochastic steady-state Tracking Error Minimization (CV-STEM), a new state feedback control framework for a class of It\u00f4 stochastic nonlinear systems and Lagrangian systems. Its innovation lies in computing the control input by an optimal contraction metric, which greedily minimizes an upper bound of the steady-state mean squared tracking error of the system trajectories. Although the problem of minimizing the bound is nonconvex, its equivalent convex formulation is proposed utilizing SDC parameterizations of the nonlinear system equation. It is shown using stochastic incremental contraction analysis that the CV-STEM provides a sufficient guarantee for exponential boundedness of the error for all time with L\u2082-robustness properties. For the sake of its sampling-based implementation, we present discrete-time stochastic contraction analysis with respect to a state- and time-dependent metric along with its explicit connection to continuous-time cases. We validate the superiority of the CV-STEM to PID, H\u221e, and baseline nonlinear controllers for spacecraft attitude control and synchronization problems.", "date": "2021-10", "date_type": "published", "publication": "IEEE Transactions on Automatic Control", "volume": "66", "number": "10", "publisher": "IEEE", "pagerange": "4731-4746", "id_number": "CaltechAUTHORS:20191029-154243537", "issn": "0018-9286", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20191029-154243537", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "JPL/Caltech" }, { "agency": "Raytheon Company" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/TAC.2020.3038402", "primary_object": { "basename": "09261103.pdf", "url": "https://authors.library.caltech.edu/records/38h8h-egc30/files/09261103.pdf" }, "related_objects": [ { "basename": "2006.04359.pdf", "url": "https://authors.library.caltech.edu/records/38h8h-egc30/files/2006.04359.pdf" } ], "pub_year": "2021", "author_list": "Tsukamoto, Hiroyasu and Chung, Soon-Jo" }, { "id": "https://authors.library.caltech.edu/records/1khj9-scp36", "eprint_id": 109046, "eprint_status": "archive", "datestamp": "2023-08-20 05:17:55", "lastmod": "2026-03-30 14:25:34", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Rivi\u00e8re-Benjamin", "name": { "family": "Rivi\u00e8re", "given": "Benjamin" }, "orcid": "0000-0003-4189-4090" }, { "id": "Hoenig-Wolfgang", "name": { "family": "Hoenig", "given": "Wolfgang" }, "orcid": "0000-0002-0773-028X" }, { "id": "Anderson-Matthew-J", "name": { "family": "Anderson", "given": "Matthew" }, "orcid": "0000-0001-8884-3448" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Neural Tree Expansion for Multi-Robot Planning in Non-Cooperative Environments", "ispublished": "pub", "full_text_status": "public", "keywords": "Distributed robot systems, motion and path planning, reinforcement learning", "note": "\u00a9 2021 IEEE. \n\nManuscript received February 24, 2021; accepted June 24, 2021. Date of publication July 14, 2021; date of current version July 26, 2021. \n\nThis work was supported by the Defense Advanced Research Projects Agency (DARPA). The views, opinions and/or findings expressed are those of the authors, and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government. Preliminary work was in part funded by Raytheon. Video: https://youtu.be/mklbTfWl7DE. Code: https://github.com/bpriviere/decision_making.\n\nSubmitted - 2104.09705.pdf
", "abstract": "We present a self-improving, Neural Tree Expansion (NTE) method for multi-robot online planning in non-cooperative environments, where each robot attempts to maximize its cumulative reward while interacting with other self-interested robots. Our algorithm adapts the centralized, perfect information, discrete-action space method from AlphaZero to a decentralized, partial information, continuous action space setting for multi-robot applications. Our method has three interacting components: (i) a centralized, perfect-information \"expert\" Monte Carlo Tree Search (MCTS) with large computation resources that provides expert demonstrations, (ii) a decentralized, partial-information \"learner\" MCTS with small computation resources that runs in real-time and provides self-play examples, and (iii) policy & value neural networks that are trained with the expert demonstrations and bias both the expert and the learner tree growth. Our numerical experiments demonstrate Neural Tree Expansion's computational advantage by finding better solutions than a MCTS with 20 times more resources. The resulting policies are dynamically sophisticated, demonstrate coordination between robots, and play the Reach-Target-Avoid differential game significantly better than the state-of-the-art control-theoretic baseline for multi-robot, double-integrator systems. Our hardware experiments on an aerial swarm demonstrate the computational advantage of Neural Tree Expansion, enabling online planning at 20 Hz with effective policies in complex scenarios.", "date": "2021-10", "date_type": "published", "publication": "IEEE Robotics and Automation Letters", "volume": "6", "number": "4", "publisher": "IEEE", "pagerange": "6868-6875", "id_number": "CaltechAUTHORS:20210510-141334067", "issn": "2377-3766", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20210510-141334067", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Defense Advanced Research Projects Agency (DARPA)" }, { "agency": "Raytheon Company" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/LRA.2021.3096758", "primary_object": { "basename": "2104.09705.pdf", "url": "https://authors.library.caltech.edu/records/1khj9-scp36/files/2104.09705.pdf" }, "pub_year": "2021", "author_list": "Rivi\u00e8re, Benjamin; Hoenig, Wolfgang; et al." }, { "id": "https://authors.library.caltech.edu/records/5zy30-t4193", "eprint_id": 103813, "eprint_status": "archive", "datestamp": "2023-08-22 10:06:45", "lastmod": "2026-03-30 15:20:57", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Matsuka-Kai", "name": { "family": "Matsuka", "given": "Kai" }, "orcid": "0000-0003-2116-9756" }, { "id": "Feldman-Aaron-O", "name": { "family": "Feldman", "given": "Aaron O." } }, { "id": "Lupu-Elena-S", "name": { "family": "Lupu", "given": "Elena S." }, "orcid": "0000-0002-3968-2630" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" }, { "id": "Hadaegh-Fred-Y", "name": { "family": "Hadaegh", "given": "Fred Y." } } ] }, "title": "Decentralized formation pose estimation for spacecraft swarms", "ispublished": "pub", "full_text_status": "public", "keywords": "Swarm localization; Spacecraft swarm; Large scale estimation; Decentralized estimation", "note": "\u00a9 2020 Published by Elsevier Ltd on behalf of COSPAR. \n\nAccepted 11 June 2020, Available online 26 June 2020. \n\nThis research was supported in part by the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. The work of Kai Matsuka was supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE 1745301. We also would like to thank Alexei Harvard for his help on camera calibration as well as to Jennifer Sun and Amir Rahmani for their technical support. \n\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\n\nAccepted Version - ASR_final.pdf
", "abstract": "For spacecraft swarms, the multi-agent localization algorithm must scale well with the number of spacecraft and adapt to time-varying communication and relative sensing networks. In this paper, we present a decentralized, scalable algorithm for swarm localization, called the Decentralized Pose Estimation (DPE) algorithm. The DPE considers both communication and relative sensing graphs and defines an observable local formation. Each spacecraft jointly localizes its local subset of spacecraft using direct and communicated measurements. Since the algorithm is local, the algorithm complexity does not grow with the number of spacecraft in the swarm. As part of the DPE, we present the Swarm Reference Frame Estimation (SRFE) algorithm, a distributed consensus algorithm to co-estimate a common Local-Vertical, Local-Horizontal (LVLH) frame. The DPE combined with the SRFE provides a scalable, fully-decentralized navigation solution that can be used for swarm control and motion planning. Numerical simulations and experiments using Caltech's robotic spacecraft simulators are presented to validate the effectiveness and scalability of the DPE algorithm.", "date": "2021-06-01", "date_type": "published", "publication": "Advances in Space Research", "volume": "67", "number": "11", "publisher": "Elsevier", "pagerange": "3527-3545", "id_number": "CaltechAUTHORS:20200610-090345377", "issn": "0273-1177", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20200610-090345377", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "NASA/JPL/Caltech" }, { "agency": "NSF Graduate Research Fellowship", "grant_number": "DGE-1745301" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1016/j.asr.2020.06.016", "primary_object": { "basename": "ASR_final.pdf", "url": "https://authors.library.caltech.edu/records/5zy30-t4193/files/ASR_final.pdf" }, "pub_year": "2021", "author_list": "Matsuka, Kai; Feldman, Aaron O.; et al." }, { "id": "https://authors.library.caltech.edu/records/m40dq-4s262", "eprint_id": 103472, "eprint_status": "archive", "datestamp": "2023-08-20 02:27:33", "lastmod": "2026-03-30 15:29:01", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Nakka-Yashwanth-K", "name": { "family": "Nakka", "given": "Yashwanth Kumar" }, "orcid": "0000-0001-7897-3644" }, { "id": "Liu-Anqi", "name": { "family": "Liu", "given": "Anqi" } }, { "id": "Shi-Guanya", "name": { "family": "Shi", "given": "Guanya" }, "orcid": "0000-0002-9075-3705" }, { "id": "Anandkumar-A", "name": { "family": "Anandkumar", "given": "Anima" } }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Chance-Constrained Trajectory Optimization for Safe Exploration and Learning of Nonlinear Systems", "ispublished": "pub", "full_text_status": "public", "note": "\u00a9 2020 IEEE. \n\nManuscript receivedMay 8, 2020; accepted October 1, 2020. Date of publication December 10, 2020; date of current version December 28, 2020. \n\nThis letter was recommended for publication by Associate Editor L. Tapia and Editor N. Amato upon evaluation of the reviewers' comments. This work was supported by the Jet Propulsion Laboratory, Caltech and the Raytheon Company. The work of Anqi Liu was supported by a PIMCO Postdoctoral Fellowship. \n\nWe acknowledge the contribution of Irene S. Crowell in\nimplementing Info-SNOC.\n\nSubmitted - 2005.04374.pdf
", "abstract": "Learning-based control algorithms require data collection with abundant supervision for training. Safe exploration algorithms ensure the safety of this data collection process even when only partial knowledge is available. We present a new approach for optimal motion planning with safe exploration that integrates chance-constrained stochastic optimal control with dynamics learning and feedback control. We derive an iterative convex optimization algorithm that solves an Information-cost Stochastic Nonlinear Optimal Control problem (Info-SNOC). The optimization objective encodes control cost for performance and exploration cost for learning, and the safety is incorporated as distributionally robust chance constraints. The dynamics are predicted from a robust regression model that is learned from data. The Info-SNOC algorithm is used to compute a sub-optimal pool of safe motion plans that aid in exploration for learning unknown residual dynamics under safety constraints. A stable feedback controller is used to execute the motion plan and collect data for model learning. We prove the safety of rollout from our exploration method and reduction in uncertainty over epochs, thereby guaranteeing the consistency of our learning method. We validate the effectiveness of Info-SNOC by designing and implementing a pool of safe trajectories for a planar robot. We demonstrate that our approach has higher success rate in ensuring safety when compared to a deterministic trajectory optimization approach.", "date": "2021-04", "date_type": "published", "publication": "IEEE Robotics and Automation Letters", "volume": "6", "number": "2", "publisher": "IEEE", "pagerange": "389-396", "id_number": "CaltechAUTHORS:20200526-150616242", "issn": "2377-3766", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20200526-150616242", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "JPL/Caltech" }, { "agency": "Raytheon Company" }, { "agency": "PIMCO" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/LRA.2020.3044033", "primary_object": { "basename": "2005.04374.pdf", "url": "https://authors.library.caltech.edu/records/m40dq-4s262/files/2005.04374.pdf" }, "pub_year": "2021", "author_list": "Nakka, Yashwanth Kumar; Liu, Anqi; et al." }, { "id": "https://authors.library.caltech.edu/records/vmjhy-kmj35", "eprint_id": 109031, "eprint_status": "archive", "datestamp": "2023-08-20 02:35:27", "lastmod": "2026-03-29 20:08:03", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Voloshin-Cameron", "name": { "family": "Voloshin", "given": "Cameron" } }, { "id": "Jiang-Nan", "name": { "family": "Jiang", "given": "Nan" } }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" } ] }, "title": "Minimax Model Learning", "ispublished": "pub", "full_text_status": "public", "note": "\u00a9 2021 by the author(s). \n\nProceedings of the 24th International Conference on Artificial Intelligence and Statistics (AISTATS) 2021, San Diego, California, USA. PMLR: Volume 130. \n\nCameron Voloshin is supported in part by a Kortschak Fellowship. This work is also supported in part by NSF # 1645832, NSF # 1918839, and funding from Beyond Limits. Nan Jiang is sponsored in part by the DEVCOM Army Research Laboratory under Cooperative Agreement W911NF-17-2-0196 (ARL IoBT CRA). The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Laboratory or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.\n\nPublished - voloshin21a.pdf
Submitted - 2103.02084.pdf
Supplemental Material - voloshin21a-supp.pdf
", "abstract": "We present a novel off-policy loss function for learning a transition model in model-based reinforcement learning. Notably, our loss is derived from the off-policy policy evaluation objective with an emphasis on correcting distribution shift. Compared to previous model-based techniques, our approach allows for greater robustness under model misspecification or distribution shift induced by learning/evaluating policies that are distinct from the data-generating policy. We provide a theoretical analysis and show empirical improvements over existing model-based off-policy evaluation methods. We provide further analysis showing our loss can be used for off-policy optimization (OPO) and demonstrate its integration with more recent improvements in OPO.", "date": "2021-04", "date_type": "published", "publication": "Proceedings of Machine Learning Research", "volume": "130", "publisher": "PMLR", "pagerange": "1612-1620", "id_number": "CaltechAUTHORS:20210510-100815979", "issn": "2640-3498", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20210510-100815979", "funders": { "items": [ { "agency": "Caltech" }, { "agency": "NSF", "grant_number": "CNS-1645832" }, { "agency": "NSF", "grant_number": "IIS-1918839" }, { "agency": "Beyond Limits" }, { "agency": "Army Research Office (ARO)", "grant_number": "W911NF-17-2-0196" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.48550/arXiv.2103.02084", "primary_object": { "basename": "2103.02084.pdf", "url": "https://authors.library.caltech.edu/records/vmjhy-kmj35/files/2103.02084.pdf" }, "related_objects": [ { "basename": "voloshin21a.pdf", "url": "https://authors.library.caltech.edu/records/vmjhy-kmj35/files/voloshin21a.pdf" }, { "basename": "voloshin21a-supp.pdf", "url": "https://authors.library.caltech.edu/records/vmjhy-kmj35/files/voloshin21a-supp.pdf" } ], "pub_year": "2021", "author_list": "Voloshin, Cameron; Jiang, Nan; et al." }, { "id": "https://authors.library.caltech.edu/records/vat05-w9c33", "eprint_id": 106577, "eprint_status": "archive", "datestamp": "2023-08-20 02:28:47", "lastmod": "2026-03-30 07:48:44", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Zhao-Eric", "name": { "family": "Zhao", "given": "Eric" }, "orcid": "0000-0002-9595-0150" }, { "id": "Liu-Anqi", "name": { "family": "Liu", "given": "Anqi" } }, { "id": "Anandkumar-A", "name": { "family": "Anandkumar", "given": "Animashree" } }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" } ] }, "title": "Active Learning under Label Shift", "ispublished": "pub", "full_text_status": "public", "note": "\u00a9 2021 by the author(s). \n\nAnqi Liu is supported by the PIMCO Postdoctoral Fellowship. Prof. Anandkumar is supported by Bren endowed Chair, faculty awards from Microsoft, Google, and Adobe, Beyond Limits, and LwLL grants. This work is also supported by funding from Raytheon and NASA TRISH.\n\nPublished - zhao21b.pdf
Submitted - 2007.08479.pdf
Supplemental Material - zhao21b-supp.pdf
", "abstract": "We address the problem of active learning under label shift: when the class proportions of source and target domains differ. We introduce a \"medial distribution\" to incorporate a tradeoff between importance weighting and class-balanced sampling and propose their combined usage in active learning. Our method is known as Mediated Active Learning under Label Shift (MALLS). It balances the bias from class-balanced sampling and the variance from importance weighting. We prove sample complexity and generalization guarantees for MALLS which show active learning reduces asymptotic sample complexity even under arbitrary label shift. We empirically demonstrate MALLS scales to high-dimensional datasets and can reduce the sample complexity of active learning by 60% in deep active learning tasks.", "date": "2021-04", "date_type": "published", "publication": "Proceedings of Machine Learning Research", "volume": "130", "publisher": "PMLR", "pagerange": "3412-3420", "id_number": "CaltechAUTHORS:20201110-074357009", "issn": "2640-3498", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20201110-074357009", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "PIMCO Postdoctoral Fellowship" }, { "agency": "Bren Professor of Computing and Mathematical Sciences" }, { "agency": "Microsoft Faculty Fellowship" }, { "agency": "Google Faculty Research Award" }, { "agency": "Adobe" }, { "agency": "Learning with Less Labels (LwLL)" }, { "agency": "Raytheon Company" }, { "agency": "NASA" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.48550/arXiv.2007.08479", "primary_object": { "basename": "2007.08479.pdf", "url": "https://authors.library.caltech.edu/records/vat05-w9c33/files/2007.08479.pdf" }, "related_objects": [ { "basename": "zhao21b.pdf", "url": "https://authors.library.caltech.edu/records/vat05-w9c33/files/zhao21b.pdf" }, { "basename": "zhao21b-supp.pdf", "url": "https://authors.library.caltech.edu/records/vat05-w9c33/files/zhao21b-supp.pdf" } ], "pub_year": "2021", "author_list": "Zhao, Eric; Liu, Anqi; et al." }, { "id": "https://authors.library.caltech.edu/records/7mf7x-pvr53", "eprint_id": 106090, "eprint_status": "archive", "datestamp": "2023-08-20 02:01:41", "lastmod": "2026-03-29 21:34:07", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Khojasteh-Mohammad-Javad", "name": { "family": "Khojasteh", "given": "Mohammad Javad" }, "orcid": "0000-0002-8459-6483" }, { "id": "Hedayatpour-Mojtaba", "name": { "family": "Hedayatpour", "given": "Mojtaba" } }, { "id": "Cort\u00e9s-Jorge", "name": { "family": "Cort\u00e9s", "given": "Jorge" }, "orcid": "0000-0001-9582-5184" }, { "id": "Franceschetti-Massimo", "name": { "family": "Franceschetti", "given": "Massimo" }, "orcid": "0000-0002-4057-8152" } ] }, "title": "Exploiting Timing Information in Event-Triggered Stabilization of Linear Systems With Disturbances", "ispublished": "pub", "full_text_status": "restricted", "keywords": "Control under communication constraints,\nevent-triggered control, feedback stabilization with delay,\nnetworked control systems", "note": "\u00a9 2020 IEEE. \n\nManuscript received February 1, 2020; revised February 3, 2020, July 21, 2020, and August 29, 2020; accepted September 12, 2020. Date of publication October 9, 2020; date of current version February 26, 2021. \n\nThis work was supported in part by NSF under Award CNS-1446891 and Award ECCS-1917177.", "abstract": "Similar to the way pauses are used in spoken language to convey information, it is also possible to transmit information in communication networks not only by message content, but also with its timing. This article presents an event-triggering strategy that utilizes timing information by transmitting in a state-dependent fashion. We consider the stabilization of a continuous-time, time-invariant, linear plant over a digital communication channel with bounded delay and subject to bounded plant disturbances, and establish two main results. On the one hand, we design an encoding\u2013decoding scheme that guarantees a sufficient information transmission rate for stabilization. On the other hand, we determine a lower bound on the information transmission rate necessary for stabilization by any control policy.", "date": "2021-03", "date_type": "published", "publication": "IEEE Transactions on Control of Network Systems", "volume": "8", "number": "1", "publisher": "IEEE", "pagerange": "15-27", "id_number": "CaltechAUTHORS:20201015-152732966", "issn": "2325-5870", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20201015-152732966", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "NSF", "grant_number": "CNS-1446891" }, { "agency": "NSF", "grant_number": "ECCS-1917177" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/tcns.2020.3030008", "pub_year": "2021", "author_list": "Khojasteh, Mohammad Javad; Hedayatpour, Mojtaba; et al." }, { "id": "https://authors.library.caltech.edu/records/c1zzq-23c35", "eprint_id": 106094, "eprint_status": "archive", "datestamp": "2023-08-22 08:36:12", "lastmod": "2026-03-30 16:21:31", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Maser-Michael-R", "name": { "family": "Maser", "given": "Michael R." } }, { "id": "Cui-Alexander-Y", "name": { "family": "Cui", "given": "Alexander Y." } }, { "id": "Ryou-Serim", "name": { "family": "Ryou", "given": "Serim" } }, { "id": "DeLano-Travis-J", "name": { "family": "DeLano", "given": "Travis J." }, "orcid": "0000-0002-2052-611X" }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" }, { "id": "Reisman-S-E", "name": { "family": "Reisman", "given": "Sarah E." }, "orcid": "0000-0001-8244-9300" } ] }, "title": "Multilabel Classification Models for the Prediction of Cross-Coupling Reaction Conditions", "ispublished": "pub", "full_text_status": "public", "keywords": "machine learning; graph neural network; graph attention; gradient-boosting machines; reaction condition prediction; cross-coupling; predictive modeling; molecular machine learning", "note": "\u00a9 2021 American Chemical Society. \n\nReceived: October 23, 2020; Publication Date: January 8, 2021. \n\nWe thank Prof Pietro Perona for mentorship guidance and helpful project discussions and Chase Blagden for help in structuring the GBM experiments. Fellowship support was provided by the NSF (M.R.M., T.J.D. Grant No. DGE-1144469). S.E.R. is a Heritage Medical Research Institute Investigator. Y.Y. is supported in part by NSF 1645832 and NSF 1918839 and funding from Raytheon and Beyond Limits. S.R. is supported by grants from Disney Research and from Nissan Corporation. Financial support from Research Corporation is warmly acknowledged. \n\nAuthor Contributions: M.R.M., A.Y.C., and S.R. contributed equally to this work. \n\nThe authors declare no competing financial interest.\n\nSubmitted - Multi-Label_Classification_Models_for_the_Prediction_of_Cross-Coupling_Reaction_Conditions_v1.pdf
Supplemental Material - ci0c01234_si_001.pdf
", "abstract": "Machine-learned ranking models have been developed for the prediction of substrate-specific cross-coupling reaction conditions. Data sets of published reactions were curated for Suzuki, Negishi, and C\u2013N couplings, as well as Pauson\u2013Khand reactions. String, descriptor, and graph encodings were tested as input representations, and models were trained to predict the set of conditions used in a reaction as a binary vector. Unique reagent dictionaries categorized by expert-crafted reaction roles were constructed for each data set, leading to context-aware predictions. We find that relational graph convolutional networks and gradient-boosting machines are very effective for this learning task, and we disclose a novel reaction-level graph attention operation in the top-performing model.", "date": "2021-01-25", "date_type": "published", "publication": "Journal of Chemical Information and Modeling", "volume": "61", "number": "1", "publisher": "American Chemical Society", "pagerange": "156-166", "id_number": "CaltechAUTHORS:20201015-152733539", "issn": "1549-9596", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20201015-152733539", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "NSF Graduate Research Fellowship", "grant_number": "DGE-1144469" }, { "agency": "Heritage Medical Research Institute" }, { "agency": "NSF", "grant_number": "CNS-1645832" }, { "agency": "NSF", "grant_number": "1918839" }, { "agency": "Raytheon Company" }, { "agency": "Beyond Limits" }, { "agency": "Disney Research" }, { "agency": "Nissan Corporation" }, { "agency": "Research Corporation" } ] }, "local_group": { "items": [ { "id": "Heritage-Medical-Research-Institute" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1021/acs.jcim.0c01234", "primary_object": { "basename": "Multi-Label_Classification_Models_for_the_Prediction_of_Cross-Coupling_Reaction_Conditions_v1.pdf", "url": "https://authors.library.caltech.edu/records/c1zzq-23c35/files/Multi-Label_Classification_Models_for_the_Prediction_of_Cross-Coupling_Reaction_Conditions_v1.pdf" }, "related_objects": [ { "basename": "ci0c01234_si_001.pdf", "url": "https://authors.library.caltech.edu/records/c1zzq-23c35/files/ci0c01234_si_001.pdf" } ], "pub_year": "2021", "author_list": "Maser, Michael R.; Cui, Alexander Y.; et al." }, { "id": "https://authors.library.caltech.edu/records/zmec9-3ky05", "eprint_id": 104021, "eprint_status": "archive", "datestamp": "2023-08-20 01:11:14", "lastmod": "2026-03-30 15:25:08", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Tsukamoto-Hiroyasu", "name": { "family": "Tsukamoto", "given": "Hiroyasu" } }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "Neural Contraction Metrics for Robust Estimation and Control: A Convex Optimization Approach", "ispublished": "pub", "full_text_status": "public", "keywords": "Machine learning, Observers for nonlinear systems, Optimal control", "note": "\u00a9 2020 The Author(s). This work is licensed under a Creative Commons Attribution 4.0 License. \n\nManuscript received March 17, 2020; revised May 15, 2020; accepted June 4, 2020. Date of publication June 11, 2020; date of current version June 24, 2020. \n\nThis work was supported in part by the Jet Propulsion Laboratory, California Institute of Technology and in part by the Raytheon Company. Recommended by Senior Editor G. Cherubini.\n\nPublished - 09115010.pdf
Submitted - 2006.04361.pdf
", "abstract": "This letter presents a new deep learning-based framework for robust nonlinear estimation and control using the concept of a Neural Contraction Metric (NCM). The NCM uses a deep long short-term memory recurrent neural network for a global approximation of an optimal contraction metric, the existence of which is a necessary and sufficient condition for exponential stability of nonlinear systems. The optimality stems from the fact that the contraction metrics sampled offline are the solutions of a convex optimization problem to minimize an upper bound of the steady-state Euclidean distance between perturbed and unperturbed system trajectories. We demonstrate how to exploit NCMs to design an online optimal estimator and controller for nonlinear systems with bounded disturbances utilizing their duality. The performance of our framework is illustrated through Lorenz oscillator state estimation and spacecraft optimal motion planning problems.", "date": "2021-01", "date_type": "published", "publication": "IEEE Control Systems Letters", "volume": "5", "number": "1", "publisher": "IEEE", "pagerange": "211-216", "id_number": "CaltechAUTHORS:20200624-155134352", "issn": "2475-1456", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20200624-155134352", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "JPL/Caltech" }, { "agency": "Raytheon Company" } ] }, "local_group": { "items": [ { "id": "GALCIT" }, { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" } ] }, "doi": "10.1109/lcsys.2020.3001646", "primary_object": { "basename": "09115010.pdf", "url": "https://authors.library.caltech.edu/records/zmec9-3ky05/files/09115010.pdf" }, "related_objects": [ { "basename": "2006.04361.pdf", "url": "https://authors.library.caltech.edu/records/zmec9-3ky05/files/2006.04361.pdf" } ], "pub_year": "2021", "author_list": "Tsukamoto, Hiroyasu and Chung, Soon-Jo" }, { "id": "https://authors.library.caltech.edu/records/wg2w8-e6v42", "eprint_id": 103207, "eprint_status": "archive", "datestamp": "2023-08-19 22:04:25", "lastmod": "2026-03-30 16:23:59", "type": "article", "metadata_visibility": "show", "creators": { "items": [ { "id": "Rivi\u00e8re-B", "name": { "family": "Rivi\u00e8re", "given": "Benjamin" }, "orcid": "0000-0003-4189-4090" }, { "id": "H\u00f6nig-W", "name": { "family": "H\u00f6nig", "given": "Wolfgang" }, "orcid": "0000-0002-0773-028X" }, { "id": "Yue-Yisong", "name": { "family": "Yue", "given": "Yisong" }, "orcid": "0000-0001-9127-1989" }, { "id": "Chung-Soon-Jo", "name": { "family": "Chung", "given": "Soon-Jo" }, "orcid": "0000-0002-6657-3907" } ] }, "title": "GLAS: Global-to-Local Safe Autonomy Synthesis for Multi-Robot Motion Planning with End-to-End Learning", "ispublished": "pub", "full_text_status": "public", "keywords": "Distributed Robot Systems, Path Planning for Multiple Mobile Robots or Agents, Imitation Learning", "note": "\u00a9 2020 IEEE. \n\nManuscript received February 24, 2020; accepted April 20, 2020. Date of publicationMay 11, 2020; date of current version May 25, 2020. \n\nThis letter was recommended for publication by Associate Editor M. Ani Hsieh and Editor N.Y. Chong upon evaluation of the reviewers' comments. This work was supported by the Raytheon Company and Caltech/NASA Jet Propulsion Laboratory. Video: https://youtu.be/z9LjSfLfG6c. Code: https://github.com/bpriviere/glas.\n\nSubmitted - 2002.11807.pdf
", "abstract": "We present GLAS: Global-to- Local Autonomy Synthesis, a provably-safe, automated distributed policy generation for multi-robot motion planning. Our approach combines the advantage of centralized planning of avoiding local minima with the advantage of decentralized controllers of scalability and distributed computation. In particular, our synthesized policies only require relative state information of nearby neighbors and obstacles, and compute a provably-safe action. Our approach has three major components: i) we generate demonstration trajectories using a global planner and extract local observations from them, ii) we use deep imitation learning to learn a decentralized policy that can run efficiently online, and iii) we introduce a novel differentiable safety module to ensure collision-free operation, thereby allowing for end-to-end policy training. Our numerical experiments demonstrate that our policies have a 20% higher success rate than optimal reciprocal collision avoidance, ORCA, across a wide range of robot and obstacle densities. We demonstrate our method on an aerial swarm, executing the policy on low-end microcontrollers in real-time.", "date": "2020-07", "date_type": "published", "publication": "IEEE Robotics and Automation Letters", "volume": "5", "number": "3", "publisher": "Institute of Electrical and Electronics Engineers", "pagerange": "4249-4256", "id_number": "CaltechAUTHORS:20200514-141356088", "issn": "2377-3766", "official_url": "https://resolver.caltech.edu/CaltechAUTHORS:20200514-141356088", "rights": "No commercial reproduction, distribution, display or performance rights in this work are provided.", "funders": { "items": [ { "agency": "Raytheon Company" }, { "agency": "NASA/JPL/Caltech" } ] }, "local_group": { "items": [ { "id": "Center-for-Autonomous-Systems-and-Technologies-(CAST)" }, { "id": "GALCIT" } ] }, "doi": "10.1109/lra.2020.2994035", "primary_object": { "basename": "2002.11807.pdf", "url": "https://authors.library.caltech.edu/records/wg2w8-e6v42/files/2002.11807.pdf" }, "pub_year": "2020", "author_list": "Rivi\u00e8re, Benjamin; H\u00f6nig, Wolfgang; et al." } ]