Stackelberg Meta-Learning for Strategic Guidance in Multi-Robot Trajectory Planning

TL;DR

This paper introduces a Stackelberg meta-learning framework enabling a leader robot to quickly adapt its trajectory guidance strategies for different followers in cooperative tasks, even with incomplete information.

Contribution

It formulates the trajectory guidance as a dynamic Stackelberg game and applies meta-learning to enable rapid adaptation to various followers, improving cooperation efficiency.

Findings

Better generalization to different followers compared to other methods

Faster adaptation to specific followers with limited data

Enhanced guidance effectiveness over zero guidance scenarios

Abstract

Trajectory guidance requires a leader robotic agent to assist a follower robotic agent to cooperatively reach the target destination. However, planning cooperation becomes difficult when the leader serves a family of different followers and has incomplete information about the followers. There is a need for learning and fast adaptation of different cooperation plans. We develop a Stackelberg meta-learning approach to address this challenge. We first formulate the guided trajectory planning problem as a dynamic Stackelberg game to capture the leader-follower interactions. Then, we leverage meta-learning to develop cooperative strategies for different followers. The leader learns a meta-best-response model from a prescribed set of followers. When a specific follower initiates a guidance query, the leader quickly adapts to the follower-specific model with a small amount of learning data and uses it to perform trajectory guidance. We use simulations to elaborate that our method provides a better generalization and adaptation performance on learning followers' behavior than other learning approaches. The value and the effectiveness of guidance are also demonstrated by the comparison with zero guidance scenarios.