Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
140 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Learning Robust Output Control Barrier Functions from Safe Expert Demonstrations (2111.09971v3)

Published 18 Nov 2021 in eess.SY, cs.LG, and cs.SY

Abstract: This paper addresses learning safe output feedback control laws from partial observations of expert demonstrations. We assume that a model of the system dynamics and a state estimator are available along with corresponding error bounds, e.g., estimated from data in practice. We first propose robust output control barrier functions (ROCBFs) as a means to guarantee safety, as defined through controlled forward invariance of a safe set. We then formulate an optimization problem to learn ROCBFs from expert demonstrations that exhibit safe system behavior, e.g., data collected from a human operator or an expert controller. When the parametrization of the ROCBF is linear, then we show that, under mild assumptions, the optimization problem is convex. Along with the optimization problem, we provide verifiable conditions in terms of the density of the data, smoothness of the system model and state estimator, and the size of the error bounds that guarantee validity of the obtained ROCBF. Towards obtaining a practical control algorithm, we propose an algorithmic implementation of our theoretical framework that accounts for assumptions made in our framework in practice. We validate our algorithm in the autonomous driving simulator CARLA and demonstrate how to learn safe control laws from simulated RGB camera images.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (57)
  1. W. Schwarting, J. Alonso-Mora, and D. Rus, “Planning and decision-making for autonomous vehicles,” An. Review Control, Robot., and Auton. Syst., vol. 1, pp. 187–210, 2018.
  2. P. Wieland and F. Allgöwer, “Constructive safety using control barrier functions,” in Proc. Symp. Nonlin. Control Syst., Pretoria, South Africa, August 2007, pp. 462–467.
  3. A. D. Ames, X. Xu, J. W. Grizzle, and P. Tabuada, “Control barrier function based quadratic programs for safety critical systems,” IEEE Trans. Autom. Control, vol. 62, no. 8, pp. 3861–3876, 2017.
  4. P. Glotfelter, J. Cortés, and M. Egerstedt, “Nonsmooth barrier functions with applications to multi-robot systems,” IEEE Control Syst. Lett., vol. 1, no. 2, pp. 310–315, 2017.
  5. L. Wang, A. D. Ames, and M. Egerstedt, “Safety barrier certificates for collisions-free multirobot systems,” IEEE Trans. Robot., vol. 33, no. 3, pp. 661–674, 2017.
  6. L. Lindemann and D. V. Dimarogonas, “Control barrier functions for signal temporal logic tasks,” IEEE Control Syst. Lett., vol. 3, no. 1, pp. 96–101, 2019.
  7. W. Xiao and C. Belta, “Control barrier functions for systems with high relative degree,” in Proc. Conf. Decis. Control, Nice, France, December 2019, pp. 474–479.
  8. S. Kolathaya and A. D. Ames, “Input-to-state safety with control barrier functions,” IEEE Control Syst. Lett., vol. 3, no. 1, pp. 108–113, 2018.
  9. X. Xu, P. Tabuada, J. W. Grizzle, and A. D. Ames, “Robustness of control barrier functions for safety critical control,” in Proc. Conf. Analys. Design Hybrid Syst., Atlanta, GA, October 2015, pp. 54–61.
  10. M. Jankovic, “Robust control barrier functions for constrained stabilization of nonlinear systems,” Automatica, vol. 96, pp. 359–367, 2018.
  11. S. Dean, A. J. Taylor, R. K. Cosner, B. Recht, and A. D. Ames, “Guaranteeing safety of learned perception modules via measurement-robust control barrier functions,” in Proc. Conf. Robot Learning, Boston, Massachusetts, November 2020, pp. 1–17.
  12. Y. Zhang, S. Walters, and X. Xu, “Control barrier function meets interval analysis: Safety-critical control with measurement and actuation uncertainties,” arXiv preprint arXiv:2110.00915, 2021.
  13. R. K. Cosner, A. W. Singletary, A. J. Taylor, T. G. Molnar, K. L. Bouman, and A. D. Ames, “Measurement-robust control barrier functions: Certainty in safety with uncertainty in state,” arXiv preprint arXiv:2104.14030, 2021.
  14. K. Garg and D. Panagou, “Robust control barrier and control lyapunov functions with fixed-time convergence guarantees,” in Proc. Am. Control Conf., New Orleans, LA, May 2021, pp. 2292–2297.
  15. P.-F. Massiani, S. Heim, F. Solowjow, and S. Trimpe, “Safe value functions,” arXiv preprint arXiv:2105.12204, 2021.
  16. J. Ferlez, M. Elnaggar, Y. Shoukry, and C. Fleming, “ShieldNN: A provably safe nn filter for unsafe nn controllers,” arXiv preprint arXiv:2006.09564, 2020.
  17. B. T. Lopez, J.-J. E. Slotine, and J. P. How, “Robust adaptive control barrier functions: An adaptive and data-driven approach to safety,” IEEE Control Syst. Lett., vol. 5, no. 3, pp. 1031–1036, 2020.
  18. A. J. Taylor and A. D. Ames, “Adaptive safety with control barrier functions,” in Proc. Am. Control Conf., Denver, CO, July 2020, pp. 1399–1405.
  19. Y. Emam, P. Glotfelter, S. Wilson, G. Notomista, and M. Egerstedt, “Data-driven robust barrier functions for safe, long-term operation,” arXiv preprint arXiv:2104.07592, 2021.
  20. A. Taylor, A. Singletary, Y. Yue, and A. Ames, “Learning for safety-critical control with control barrier functions,” in Proc. Conf. Learning Dynamics Control, San Francisco, CA, June 2020, pp. 708–717.
  21. N. Csomay-Shanklin, R. K. Cosner, M. Dai, A. J. Taylor, and A. D. Ames, “Episodic learning for safe bipedal locomotion with control barrier functions and projection-to-state safety,” in Proc. Conf. Learning Dynamics Control, Zurich, Switzerland, June 2021, pp. 1041–1053.
  22. H. Yin, P. Seiler, M. Jin, and M. Arcak, “Imitation learning with stability and safety guarantees,” IEEE Control Syst. Lett., 2021.
  23. R. Cheng, G. Orosz, R. M. Murray, and J. W. Burdick, “End-to-end safe reinforcement learning through barrier functions for safety-critical continuous control tasks,” in Proc. Conf. Artificial Intel., Honolulu, HI, February 2019, pp. 3387–3395.
  24. L. Wang, E. A. Theodorou, and M. Egerstedt, “Safe learning of quadrotor dynamics using barrier certificates,” in Proc. Conf. Robot. Automat., Brisbane, Australia, May 2018, pp. 2460–2465.
  25. C. K. Verginis, F. Djeumou, and U. Topcu, “Safety-constrained learning and control using scarce data and reciprocal barriers,” arXiv preprint arXiv:2105.06526, 2022.
  26. W. S. Cortez and D. V. Dimarogonas, “Correct-by-design control barrier functions for euler-lagrange systems with input constraints,” in Proc. Am. Control Conf., Denver, CO, July 2020, pp. 950–955.
  27. S. Prajna, A. Jadbabaie, and G. J. Pappas, “A framework for worst-case and stochastic safety verification using barrier certificates,” IEEE Trans. Autom. Control, vol. 52, no. 8, pp. 1415–1428, 2007.
  28. X. Xu, J. W. Grizzle, P. Tabuada, and A. D. Ames, “Correctness guarantees for the composition of lane keeping and adaptive cruise control,” IEEE Trans. Autom. Sci. Eng., vol. 15, no. 3, pp. 1216–1229, 2017.
  29. L. Wang, D. Han, and M. Egerstedt, “Permissive barrier certificates for safe stabilization using sum-of-squares,” in Proc. Am. Control Conf., Milwaukee, WI, June 2018, pp. 585–590.
  30. A. D. Ames, S. Coogan, M. Egerstedt, G. Notomista, K. Sreenath, and P. Tabuada, “Control barrier functions: Theory and applications,” in Proc. European Control Conf., Naples, Italy, June 2019, pp. 3420–3431.
  31. A. Clark, “Verification and synthesis of control barrier functions,” arXiv preprint arXiv:2104.14001, 2021.
  32. M. Srinivasan, A. Dabholkar, S. Coogan, and P. A. Vela, “Synthesis of control barrier functions using a supervised machine learning approach,” in Proc. Conf. Intel. Robots Syst., Las Vegas, NV, October 2020, pp. 7139–7145.
  33. M. Saveriano and D. Lee, “Learning barrier functions for constrained motion planning with dynamical systems,” in Proc. Conf. Intel. Robot Syst., Macau, China, November 2019, pp. 112–119.
  34. M. Ohnishi, G. Notomista, M. Sugiyama, and M. Egerstedt, “Constraint learning for control tasks with limited duration barrier functions,” Automatica, vol. 127, p. 109504, 2021.
  35. K. Long, C. Qian, J. Cortés, and N. Atanasov, “Learning barrier functions with memory for robust safe navigation,” IEEE Robot. Autom. Lett., vol. 6, no. 3, pp. 4931–4938, 2021.
  36. S. Yaghoubi, G. Fainekos, and S. Sankaranarayanan, “Training neural network controllers using control barrier functions in the presence of disturbances,” in Proc. Conf. Intel. Transp. Syst., Rhodes, Greece, September 2020, pp. 1–6.
  37. W. Xiao, C. A. Belta, and C. G. Cassandras, “Feasibility-guided learning for constrained optimal control problems,” in Proc. Conf. Decis. Control, Jeju Islands, South Korea, December 2020, pp. 1896–1901.
  38. S. Chen, M. Fazlyab, M. Morari, G. J. Pappas, and V. M. Preciado, “Learning lyapunov functions for hybrid systems,” in Proc. Conf. Hybrid Syst.: Comp. Control, Nashville, TN, May 2021, pp. 1–11.
  39. A. Abate, D. Ahmed, A. Edwards, M. Giacobbe, and A. Peruffo, “FOSSIL: a software tool for the formal synthesis of lyapunov functions and barrier certificates using neural networks,” in Proc. Conf. Hybrid Syst.: Comp. Control, Nashville, TN, May 2021, pp. 1–11.
  40. H. Dai, B. Landry, M. Pavone, and R. Tedrake, “Counter-example guided synthesis of neural network lyapunov functions for piecewise linear systems,” in Conf. Decis. Control, Jeju Islands, South Korea, December 2020, pp. 1274–1281.
  41. J. Kapinski, J. V. Deshmukh, S. Sankaranarayanan, and N. Arechiga, “Simulation-guided lyapunov analysis for hybrid dynamical systems,” in Proceedings of the 17th international conference on Hybrid systems: computation and control, 2014, pp. 133–142.
  42. N. M. Boffi, S. Tu, N. Matni, J.-J. E. Slotine, and V. Sindhwani, “Learning stability certificates from data,” in Proc. Conf. Robot Learning, Boston, Massachusetts, November 2020.
  43. W. Jin, Z. Wang, Z. Yang, and S. Mou, “Neural certificates for safe control policies,” arXiv preprint arXiv:2006.08465, 2020.
  44. H. Zhao, X. Zeng, T. Chen, Z. Liu, and J. Woodcock, “Learning safe neural network controllers with barrier certificates,” in Proc. Symp. Depend. Software Eng.: Theories, Tools, Appl., Guangzhou, China, November 2020, pp. 177–185.
  45. D. Sun, S. Jha, and C. Fan, “Learning certified control using contraction metric,” arXiv preprint arXiv:2011.12569, 2020.
  46. S. M. Khansari-Zadeh and A. Billard, “Learning control lyapunov function to ensure stability of dynamical system-based robot reaching motions,” Robot. Autonom. Syst., vol. 62, no. 6, pp. 752–765, 2014.
  47. A. Robey, H. Hu, L. Lindemann, H. Zhang, D. V. Dimarogonas, S. Tu, and N. Matni, “Learning control barrier functions from expert demonstrations,” in Proc. Conf. Decis. Control, Jeju Islands, South Korea, December 2020, pp. 3717–3724.
  48. A. Robey, L. Lindemann, S. Tu, and N. Matni, “Learning robust hybrid control barrier functions for uncertain systems,” in Proc. Conf. Anal. Design Hybrid Syst., Brussels, Belgium, July 2021, pp. 1–6.
  49. L. Lindemann, H. Hu, A. Robey, H. Zhang, D. V. Dimarogonas, S. Tu, and N. Matni, “Learning hybrid control barrier functions from data,” in Proc. Conf. Robot Learning, Boston, Massachusetts, November 2020.
  50. A. Dosovitskiy, G. Ros, F. Codevilla, A. Lopez, and V. Koltun, “Carla: An open urban driving simulator,” in Proc. Conf. Robot Learning, Mountain View, California, November 2017, pp. 1–16.
  51. L. Ljung, “System identification,” in Signal analysis and prediction.   Springer, 1998, pp. 163–173.
  52. S. Dean, N. Matni, B. Recht, and V. Ye, “Robust guarantees for perception-based control,” in Learning for Dynamics and Control.   PMLR, 2020, pp. 350–360.
  53. G. Chou, N. Ozay, and D. Berenson, “Safe output feedback motion planning from images via learned perception modules and contraction theory,” arXiv preprint arXiv:2206.06553, 2022.
  54. J. Köhler, M. A. Müller, and F. Allgöwer, “Robust output feedback model predictive control using online estimation bounds,” arXiv preprint arXiv:2105.03427, 2021.
  55. L. Chamon and A. Ribeiro, “Probably approximately correct constrained learning,” Advances in Neural Information Processing Systems, vol. 33, pp. 16 722–16 735, 2020.
  56. C. Xia, W. Hsu, M.-L. Lee, and B. C. Ooi, “Border: Efficient computation of boundary points,” IEEE Trans. Knowledge Data Eng., vol. 18, no. 3, pp. 289–303, 2006.
  57. A. Rahimi and B. Recht, “Random features for large-scale kernel machines,” in Proc. Advances Neur. Inform. Proc. Syst., Vancouver, Canada, December 2008, pp. 1177–1184.
Citations (28)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
Github Logo Streamline Icon: https://streamlinehq.com

GitHub

  1. GitHub - unstable-zeros/learning-rocbfs (8 stars)