Neural Lyapunov Differentiable Predictive Control
arXiv:2205.10728 · doi:10.1109/CDC51059.2022.9992386
Abstract
We present a learning-based predictive control methodology using the differentiable programming framework with probabilistic Lyapunov-based stability guarantees. The neural Lyapunov differentiable predictive control (NLDPC) learns the policy by constructing a computational graph encompassing the system dynamics, state and input constraints, and the necessary Lyapunov certification constraints, and thereafter using the automatic differentiation to update the neural policy parameters. In conjunction, our approach jointly learns a Lyapunov function that certifies the regions of state-space with stable dynamics. We also provide a sampling-based statistical guarantee for the training of NLDPC from the distribution of initial conditions. Our offline training approach provides a computationally efficient and scalable alternative to classical explicit model predictive control solutions. We substantiate the advantages of the proposed approach with simulations to stabilize the double integrator model and on an example of controlling an aircraft model.
arXiv comment8 pages; 9 figures
References in corpus (13)
- Decoupled Weight Decay Regularization
- Safe Reinforcement Learning Using Robust MPC
- Efficient representation and approximation of model predictive control laws via deep learning
- Learning an Approximate Model Predictive Controller with Guarantees
- Differentiable MPC for End-to-end Planning and Control
- Neural Lyapunov Control
- Probabilistic performance validation of deep learning-based robust NMPC controllers
- Learning Stable Deep Dynamics Models
- MPC-Inspired Neural Network Policies for Sequential Decision Making
- Learning Constrained Adaptive Differentiable Predictive Control Policies With Guarantees
- Large Scale Model Predictive Control with Neural Networks and Primal Active Sets
- Learning Stochastic Parametric Differentiable Predictive Control Policies
- Neural Identification for Control