Distance Metric Learning for RRT-Based Motion Planning for Wheeled Mobile Robots

The distance metric is a key component in RRT-based motion planning that deeply affects coverage of the state space, path quality and planning time. With the goal to speed up planning time, we introduce a learning approach to compute the distance metric for RRT-based planners. By exploiting a novel extend function which solves the two-point boundary value problem for wheeled mobile robots, we train a nonlinear parametric model that achieves good performances in regression and ranking tests. We compare our approach to three baselines and show that the new approach is faster by several factors with negligible loss of path quality. We also show that the approach covers the state space in the same way than the true distance metric.