Motion Planning Test Suite

Motion planning is the problem of finding a valid path for a robot from a start to goal configuration. In addition to robotics, it has applications to other fields in computer science such as CAD software, autonomous vehicles, and artificial intelligence. Motion planning in robotics is difficult because of the large number of variables needed to account for a robots shape, size, dynamic environments, and forces such as gravity and friction. Weve designed a test suite to evaluate motion planning algorithms, that is integrated with Parasol Lab’s C++ motion planning library (PMPL), visualization tool (Vizmo++), and a physically based robot simulator. In particular, weve developed benchmark scenarios to exercise motion planning algorithms in a variety of scenarios and for a range of robots. A representative set of problems were designed such as freebody scenarios that varied from two to three dimensions, a crane-like, robot with joints tasked with moving its arm from one location to another, and a car-like robot with different actuators that limited its movement to seem more car-like. These benchmark scenarios will enable the comparison and evaluation of future motion planning algorithms