Learning to Control under Uncertainty with Data-Based Iterative Linear Quadratic Regulator

被引:0
作者
Wang, Ran [1 ]
Goyal, Raman [1 ]
Chakravorty, Suman [1 ]
机构
[1] Texas A&M Univ, Dept Aerosp Engn, College Stn, TX 77843 USA
来源
2023 62ND IEEE CONFERENCE ON DECISION AND CONTROL, CDC | 2023年
关键词
Learning under noise; partial-state observation; data-based control; robotic motion planning;
D O I
10.1109/CDC49753.2023.10384069
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
This paper studies the learning-to-control problem under process and sensing uncertainties for dynamical systems. In our previous work, we developed a data-based generalization of the iterative linear quadratic regulator (iLQR) to design closed-loop feedback control for high-dimensional dynamical systems with partial state observation. This method required perfect simulation rollouts which are not realistic in real applications. In this work, we briefly introduce this method and explore its efficacy under process and sensing uncertainties. We prove that in the fully observed case where the system dynamics are corrupted with noise but the measurements are perfect, it still converges to the global minimum. However, in the partially observed case where both process and measurement noise exist in the system, this method converges to a biased "optimum". Thus multiple rollouts need to be averaged to retrieve the true optimum. The analysis is verified in two nonlinear robotic examples simulated in the above cases.
引用
收藏
页码:789 / 794
页数:6
相关论文
共 19 条
  • [1] Bertsekas D., 2012, DYNAMIC PROGRAMMING, V1
  • [2] Bertsekas D., 1996, NEURO DYNAMIC PROGRA
  • [3] Goyal R, 2023, Arxiv, DOI arXiv:2107.08086
  • [4] How to train your robot with deep reinforcement learning: lessons we have learned
    Ibarz, Julian
    Tan, Jie
    Finn, Chelsea
    Kalakrishnan, Mrinal
    Pastor, Peter
    Levine, Sergey
    [J]. INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH, 2021, 40 (4-5) : 698 - 721
  • [5] Sim-to-Real via Sim-to-Sim: Data-efficient Robotic Grasping via Randomized-to-Canonical Adaptation Networks
    James, Stephen
    Wohlhart, Paul
    Kalakrishnan, Mrinal
    Kalashnikov, Dmitry
    Irpan, Alex
    Ibarz, Julian
    Levine, Sergey
    Hadsell, Raia
    Bousmalis, Konstantinos
    [J]. 2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 12819 - 12629
  • [6] Kalashnikov D., 2018, ARXIV180610293, P651
  • [7] Khadka Shauharda, 2019, International conference on machine learning, V97, P3341
  • [8] Levine S, 2016, J MACH LEARN RES, V17
  • [9] Reinforcement Learning for Robust Parameterized Locomotion Control of Bipedal Robots
    Li, Zhongyu
    Cheng, Xuxin
    Peng, Xue Bin
    Abbeel, Pieter
    Levine, Sergey
    Berseth, Glen
    Sreenath, Koushil
    [J]. 2021 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2021), 2021, : 2811 - 2817
  • [10] Human-level control through deep reinforcement learning
    Mnih, Volodymyr
    Kavukcuoglu, Koray
    Silver, David
    Rusu, Andrei A.
    Veness, Joel
    Bellemare, Marc G.
    Graves, Alex
    Riedmiller, Martin
    Fidjeland, Andreas K.
    Ostrovski, Georg
    Petersen, Stig
    Beattie, Charles
    Sadik, Amir
    Antonoglou, Ioannis
    King, Helen
    Kumaran, Dharshan
    Wierstra, Daan
    Legg, Shane
    Hassabis, Demis
    [J]. NATURE, 2015, 518 (7540) : 529 - 533