Abstract:Learning-based neural network (NN) control policies have shown impressive empirical performance in a wide range of tasks in robotics and control. However, formal (Lyapunov) stability guarantees over the region-of-attraction (ROA) for NN controllers with nonlinear dynamical systems are challenging to obtain, and most existing approaches rely on expensive solvers such as sums-of-squares (SOS), mixed-integer programming (MIP), or satisfiability modulo theories (SMT). In this paper, we demonstrate a new framework for learning NN controllers together with Lyapunov certificates using fast empirical falsification and strategic regularizations. We propose a novel formulation that defines a larger verifiable region-of-attraction (ROA) than shown in the literature, and refines the conventional restrictive constraints on Lyapunov derivatives to focus only on certifiable ROAs. The Lyapunov condition is rigorously verified post-hoc using branch-and-bound with scalable linear bound propagation-based NN verification techniques. The approach is efficient and flexible, and the full training and verification procedure is accelerated on GPUs without relying on expensive solvers for SOS, MIP, nor SMT. The flexibility and efficiency of our framework allow us to demonstrate Lyapunov-stable output feedback control with synthesized NN-based controllers and NN-based observers with formal stability guarantees, for the first time in literature. Source code at <a class="link-external link-https" href="https://github.com/Verified-Intelligence/Lyapunov_Stable_NN_Controllers" rel="external noopener nofollow">this https URL</a>

Compositionally Verifiable Vector Neural Lyapunov Functions for Stability Analysis of Interconnected Nonlinear Systems

Physics-Informed Neural Network Lyapunov Functions: PDE Characterization, Learning, and Verification

Formally Verified Physics-Informed Neural Control Lyapunov Functions

Towards Learning and Verifying Maximal Neural Lyapunov Functions

Neural Network-Based Lyapunov Stability Verification for Continuous-Time Nonlinear Systems Via Mixed Integer Linear Programming

Lyapunov function construction for nonlinear stochastic dynamical systems

LyZNet: A Lightweight Python Tool for Learning and Verifying Neural Lyapunov Functions and Regions of Attraction

Model-Free Verification for Neural Network Controlled Systems

Learning Lyapunov Functions for Piecewise Affine Systems with Neural Network Controllers

Stability Analysis of Switched Linear Systems with Neural Lyapunov Functions

Neural Lyapunov Control of Unknown Nonlinear Systems with Stability Guarantees

Stability Analysis Using Quadratic Constraints for Systems With Neural Network Controllers

Interval Reachability of Nonlinear Dynamical Systems with Neural Network Controllers

A Semi-Algebraic Approach for the Computation of Lyapunov Functions

Function-dependent Neural-Network-driven State Feedback Control and Self-Verification Stability for Discrete-Time Nonlinear System

Local Stabilizing Control for Polynomial Nonlinear Systems

Neural Lyapunov Control for Interconnected Systems with Unknown Dynamics

Compositional Neural Certificates for Networked Dynamical Systems

Certificates of Nonexistence for Lyapunov-Based Stability, Stabilizability and Detectability of LPV Systems

Safety Verification of Neural Feedback Systems Based on Constrained Zonotopes

Lyapunov-stable Neural Control for State and Output Feedback: A Novel Formulation