Learning Robot Navigation Behaviors by Demonstration Using a RRT∗ Planner

This paper presents an approach for learning robot navigation behaviors from demonstration using Optimal Rapidly-exploring Random Trees (RRT∗) as main planner. A new learning algorithm combining both Inverse Reinforcement Learning (IRL) and RRT∗ is developed in order to learn the RRT∗’s cost function from demonstrations. This cost function can be used later in a regular RRT∗ for robot planning including the learned behaviors in different scenarios. Simulations show how the method is able to recover the behavior from the demonstrations.