Rapidly exploring learning trees

摘要

Inverse Reinforcement Learning (IRL) for path planning enables robots to learn cost functions for difficult tasks from demonstration, instead of hard-coding them. However, IRL methods face practical limitations that stem from the need to repeat costly planning procedures. In this paper, we propose Rapidly Exploring Learning Trees (RLT*), which learns the cost functions of Optimal Rapidly Exploring Random Trees (RRT*) from demonstration, thereby making inverse learning methods applicable to more complex tasks. Our approach extends Maximum Margin Planning to work with RRT* cost functions. Furthermore, we propose a caching scheme that greatly reduces the computational cost of this approach. Experimental results on simulated and real-robot data from a social navigation scenario show that RLT* achieves better performance at lower computational cost than existing methods. We also successfully deploy control policies learned with RLT* on a real telepresence robot.