A Generalizable Physics-informed Learning Framework for Risk Probability Estimation

被引:0
|
作者
Wang, Zhuoyuan [1 ]
Nakahira, Yorie [1 ]
机构
[1] Carnegie Mellon Univ, Elect & Comp Engn Dept, Pittsburgh, PA 15213 USA
来源
LEARNING FOR DYNAMICS AND CONTROL CONFERENCE, VOL 211 | 2023年 / 211卷
关键词
Stochastic safe control; physics-informed learning; risk probability estimation; SAFE EXPLORATION; REACHABILITY;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Accurate estimates of long-term risk probabilities and their gradients are critical for many stochastic safe control methods. However, computing such risk probabilities in real-time and in unseen or changing environments is challenging. Monte Carlo (MC) methods cannot accurately evaluate the probabilities and their gradients as an infinitesimal devisor can amplify the sampling noise. In this paper, we develop an efficient method to evaluate the probabilities of long-term risk and their gradients. The proposed method exploits the fact that long-term risk probability satisfies certain partial differential equations (PDEs), which characterize the neighboring relations between the probabilities, to integrate MC methods and physics-informed neural networks. We provide theoretical guarantees of the estimation error given certain choices of training configurations. Numerical results show the proposed method has better sample efficiency, generalizes well to unseen regions, and can adapt to systems with changing parameters. The proposed method can also accurately estimate the gradients of risk probabilities, which enables first- and second-order techniques on risk probabilities to be used for learning and control.
引用
收藏
页数:13
相关论文
共 39 条
  • [31] Reconstructing the pressure field around swimming fish using a physics-informed neural network
    Calicchia, Michael A.
    Mittal, Rajat
    Seo, Jung-Hee
    Ni, Rui
    JOURNAL OF EXPERIMENTAL BIOLOGY, 2023, 226 (08):
  • [32] Physical question, virtual answer: Optimized real-time physical simulations and physics-informed learning approaches for cargo loading stability
    Mazur, Philipp Gabriel
    Melsbach, Johannes Werner
    Schoder, Detlef
    OPERATIONS RESEARCH PERSPECTIVES, 2025, 14
  • [33] Physics -informed Learning and Control of Nonlinear Systems in Quasi-LPV Framework: An Extended Lagrangian-based Approach
    Pagar, Nikhil
    Kelkar, Atul
    Velni, Javad Mohammadpour
    IFAC PAPERSONLINE, 2024, 58 (28): : 102 - 107
  • [34] VT-PINN:Variable transformation improves physics-informed neural networks for approximating partial differential equations
    Zheng, Jiachun
    Yang, Yunlei
    APPLIED SOFT COMPUTING, 2024, 167
  • [35] On-line measurement of COD and nitrate in water against stochastic background interference based on ultraviolet-visible spectroscopy and physics-informed multi-task learning
    Liu, Jiacheng
    Yu, Tao
    Wang, Xueji
    Liu, Xiao
    Wu, Lichao
    Liu, Hong
    Zhao, Yubo
    Zhou, Guangya
    Yu, Weixing
    Hu, Bingliang
    SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY, 2024, 323
  • [36] Development and Implementation of Physics-Informed Neural ODE for Dynamics Modeling of a Fixed-Wing Aircraft Under Icing/Fault
    Ma, Jinyi
    Li, Yiyang
    Tu, Jingqi
    Zhang, Yiming
    Ai, Jianliang
    Dong, Yiqun
    GUIDANCE NAVIGATION AND CONTROL, 2024, 04 (01)
  • [37] Lyapunov-Based Physics-Informed Long Short-Term Memory (LSTM) Neural Network-Based Adaptive Control
    Hart R.G.
    Griffis E.J.
    Patil O.S.
    Dixon W.E.
    IEEE Control Systems Letters, 2024, 8 : 13 - 18
  • [38] SOLVING INVERSE STOCHASTIC PROBLEMS FROM DISCRETE PARTICLE OBSERVATIONS USING THE FOKKER-PLANCK EQUATION AND PHYSICS-INFORMED NEURAL NETWORKS
    Chen, Xiaoli
    Yang, Liu
    Duan, Jinqiao
    Karniadakis, George E. M.
    SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2021, 43 (03): : B811 - B830
  • [39] Nonlinear seismic inversion by physics-informed Caianiello convolutional neural networks for overpressure prediction of source rocks in the offshore Xihu depression, East China
    Cheng, Yifan
    Fu, Li-Yun
    JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING, 2022, 215