Benchmarking Motion Planning Algorithms: An Extensible Infrastructure for Analysis and Visualization

Motion planning is a key problem in robotics that is concerned with finding a path that satisfies a goal specification subject to constraints. In its simplest form, the solution to this problem consists of finding a path connecting two states, and the only constraint is to avoid collisions. Even for this version of the motion planning problem, there is no efficient solution for the general case [1]. The addition of differential constraints on robot motion or more general goal specifications makes motion planning even harder. Given its complexity, most planning algorithms forego completeness and optimality for slightly weaker notions such as resolution completeness, probabilistic completeness [2], and asymptotic optimality.