Generating Robust Supervision for Learning-Based Visual Navigation Using Hamilton-Jacobi Reachability

Anjian Li, Somil Bansal, Georgios Giovanis, Varun Tolani, Claire Tomlin, Mo Chen
Proceedings of the 2nd Conference on Learning for Dynamics and Control, PMLR 120:500-510, 2020.

Abstract

In Bansal et al. (2019), a novel visual navigation framework that combines learning-based and model-based approaches has been proposed. Specifically, a Convolutional Neural Network (CNN) predicts a waypoint that is used by the dynamics model for planning and tracking a trajectory to the waypoint. However, the CNN inevitably makes prediction errors, ultimately leading to collisions, especially when the robot is navigating through cluttered and tight spaces. In this paper, we present a novel Hamilton-Jacobi (HJ) reachability-based method to generate supervision for the CNN for waypoint prediction. By modeling the prediction error of the CNN as disturbances in dynamics, the proposed method generates waypoints that are robust to these disturbances, and consequently to the prediction errors. Moreover, using globally optimal HJ reachability analysis leads to predicting waypoints that are time-efficient and do not exhibit greedy behavior. Through simulations and experiments on a hardware testbed, we demonstrate the advantages of the proposed approach for navigation tasks where the robot needs to navigate through cluttered, narrow indoor environments.

Cite this Paper

BibTeX
@InProceedings{pmlr-v120-li20a, title = {Generating Robust Supervision for Learning-Based Visual Navigation Using Hamilton-Jacobi Reachability}, author = {Li, Anjian and Bansal, Somil and Giovanis, Georgios and Tolani, Varun and Tomlin, Claire and Chen, Mo}, booktitle = {Proceedings of the 2nd Conference on Learning for Dynamics and Control}, pages = {500--510}, year = {2020}, editor = {Bayen, Alexandre M. and Jadbabaie, Ali and Pappas, George and Parrilo, Pablo A. and Recht, Benjamin and Tomlin, Claire and Zeilinger, Melanie}, volume = {120}, series = {Proceedings of Machine Learning Research}, month = {10--11 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v120/li20a/li20a.pdf}, url = {https://proceedings.mlr.press/v120/li20a.html}, abstract = {In Bansal et al. (2019), a novel visual navigation framework that combines learning-based and model-based approaches has been proposed. Specifically, a Convolutional Neural Network (CNN) predicts a waypoint that is used by the dynamics model for planning and tracking a trajectory to the waypoint. However, the CNN inevitably makes prediction errors, ultimately leading to collisions, especially when the robot is navigating through cluttered and tight spaces. In this paper, we present a novel Hamilton-Jacobi (HJ) reachability-based method to generate supervision for the CNN for waypoint prediction. By modeling the prediction error of the CNN as disturbances in dynamics, the proposed method generates waypoints that are robust to these disturbances, and consequently to the prediction errors. Moreover, using globally optimal HJ reachability analysis leads to predicting waypoints that are time-efficient and do not exhibit greedy behavior. Through simulations and experiments on a hardware testbed, we demonstrate the advantages of the proposed approach for navigation tasks where the robot needs to navigate through cluttered, narrow indoor environments.} }
Endnote
%0 Conference Paper %T Generating Robust Supervision for Learning-Based Visual Navigation Using Hamilton-Jacobi Reachability %A Anjian Li %A Somil Bansal %A Georgios Giovanis %A Varun Tolani %A Claire Tomlin %A Mo Chen %B Proceedings of the 2nd Conference on Learning for Dynamics and Control %C Proceedings of Machine Learning Research %D 2020 %E Alexandre M. Bayen %E Ali Jadbabaie %E George Pappas %E Pablo A. Parrilo %E Benjamin Recht %E Claire Tomlin %E Melanie Zeilinger %F pmlr-v120-li20a %I PMLR %P 500--510 %U https://proceedings.mlr.press/v120/li20a.html %V 120 %X In Bansal et al. (2019), a novel visual navigation framework that combines learning-based and model-based approaches has been proposed. Specifically, a Convolutional Neural Network (CNN) predicts a waypoint that is used by the dynamics model for planning and tracking a trajectory to the waypoint. However, the CNN inevitably makes prediction errors, ultimately leading to collisions, especially when the robot is navigating through cluttered and tight spaces. In this paper, we present a novel Hamilton-Jacobi (HJ) reachability-based method to generate supervision for the CNN for waypoint prediction. By modeling the prediction error of the CNN as disturbances in dynamics, the proposed method generates waypoints that are robust to these disturbances, and consequently to the prediction errors. Moreover, using globally optimal HJ reachability analysis leads to predicting waypoints that are time-efficient and do not exhibit greedy behavior. Through simulations and experiments on a hardware testbed, we demonstrate the advantages of the proposed approach for navigation tasks where the robot needs to navigate through cluttered, narrow indoor environments.
APA
Li, A., Bansal, S., Giovanis, G., Tolani, V., Tomlin, C. & Chen, M.. (2020). Generating Robust Supervision for Learning-Based Visual Navigation Using Hamilton-Jacobi Reachability. Proceedings of the 2nd Conference on Learning for Dynamics and Control, in Proceedings of Machine Learning Research 120:500-510 Available from https://proceedings.mlr.press/v120/li20a.html.

Related Material