Real-Time Motion Planning and Safe Navigation in Dynamic Multi-Robot Environments

All mobile robots share the need to navigate, creating the problem of motion planning. In multi-robot domains with agents acting in parallel, highly complex and unpredictable dynamics can arise. This leads to the need for navigation calculations to be carried out within tight time constraints, so that they can be applied before the dynamics of the environment make the calculated answer obsolete. At the same time, we want the robots to navigate robustly and operate safely without collisions. While motion planning has been used for high level robot navigation, or limited to semi-static or single-robot domains, it has often been dismissed for the real-time low-level control of agents due to the limited computational time and the unpredictable dynamics. Many robots now rely on local reactive methods for immediate control of the robot, but if the reason for avoiding motion planning is execution speed, the answer is to find planners that can meet this requirement. Recent advances in traditional path planning algorithms may offer hope in resolving this type of scalability, if they can be adapted to deal with the specific problems and constraints mobile robots face. Also, in order to maintain safety, new scalable methods for maintaining collision avoidance among multiple robots are needed in order to free motion planners from the “curse of dimensionality” when considering the safety of multiple robots with realistic physical dynamics constraints. This thesis contributes the pairing of real-time motion planning which builds on existing modern path planners, and a novel cooperative dynamics safety algorithm for high speed navigation of multiple agents in dynamic domains. It also explores near real-time kinematically limited motion planning for more complex environments. The thesis algorithms have been fully implemented and tested with success on multiple real robot platforms.