Chance-Constrained Iterative Linear-Quadratic Stochastic Games

TL;DR

This paper introduces CCILQGames, an algorithm for solving chance-constrained stochastic games in multi-robot planning, ensuring safety under uncertainty through an augmented Lagrangian approach.

Contribution

It presents a novel chance-constrained iterative LQ stochastic game algorithm that effectively handles safety constraints without trial-and-error penalty tuning.

Findings

Successfully applied to autonomous driving scenarios.

Generates safe, interactive strategies in stochastic environments.

Validated through experiments and Monte Carlo tests.

Abstract

Dynamic game arises as a powerful paradigm for multi-robot planning, for which safety constraint satisfaction is crucial. Constrained stochastic games are of particular interest, as real-world robots need to operate and satisfy constraints under uncertainty. Existing methods for solving stochastic games handle chance constraints using exponential penalties with hand-tuned weights. However, finding a suitable penalty weight is nontrivial and requires trial and error. In this paper, we propose the chance-constrained iterative linear-quadratic stochastic games (CCILQGames) algorithm. CCILQGames solves chance-constrained stochastic games using the augmented Lagrangian method. We evaluate our algorithm in three autonomous driving scenarios, including merge, intersection, and roundabout. Experimental results and Monte Carlo tests show that CCILQGames can generate safe and interactive strategies in stochastic environments.