Rapidly exploring learning trees
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 R...
| Main Authors: | , , |
|---|---|
| Format: | Conference item |
| Published: |
IEEE
2017
|
| Summary: | 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. |
|---|