Guaranteed Reach-Avoid for Black-Box Systems through Narrow Gaps via Neural Network Reachability

TL;DR

NeuralPARC is a novel method that guarantees safe reach-avoid performance for black-box systems modeled by neural networks, effectively handling narrow gaps and uncertainties in real-world autonomous vehicle scenarios.

Contribution

The paper introduces NeuralPARC, extending reachability analysis to neural network models of black-box systems, providing safety guarantees in complex navigation tasks.

Findings

Outperforms previous PARC method in safety and efficiency.

Successfully applied to real-world vehicle parking and autonomous surface vehicle control.

Enables provably-safe maneuvers despite large disturbances and model uncertainties.

Abstract

In the classical reach-avoid problem, autonomous mobile robots are tasked to reach a goal while avoiding obstacles. However, it is difficult to provide guarantees on the robot's performance when the obstacles form a narrow gap and the robot is a black-box (i.e. the dynamics are not known analytically, but interacting with the system is cheap). To address this challenge, this paper presents NeuralPARC. The method extends the authors' prior Piecewise Affine Reach-avoid Computation (PARC) method to systems modeled by rectified linear unit (ReLU) neural networks, which are trained to represent parameterized trajectory data demonstrated by the robot. NeuralPARC computes the reachable set of the network while accounting for modeling error, and returns a set of states and parameters with which the black-box system is guaranteed to reach the goal and avoid obstacles. NeuralPARC is shown to outperform PARC, generating provably-safe extreme vehicle drift parking maneuvers in simulations and in real life on a model car, as well as enabling safety on an autonomous surface vehicle (ASV) subjected to large disturbances and controlled by a deep reinforcement learning (RL) policy.