A practical framework for formation planning and control of multiple unmanned ground vehicles

Finding trajectories in real-time for multiple vehicles moving in a dynamic environment while satisfying all constraints is challenging. For some special groups of systems, group flatness and formation laws allow for a parameterization of the member's trajectories by the leader's flat outputs, which can significantly reduce the problem's complexity.; In the literature, most studies have focused on one aspect of the robot motion planning and control problem. Robot path planning is typically addressed using artificial intelligence techniques. Robot motion coordination and control schemes are developed using control techniques. The path planning and motion coordination and control are usually treated as two separate problems. In this research, we bridge the gap between path planning, trajectory generation, and motion control. A technique for real-time path planning has been developed and implemented. A practical framework for on-line planning and control of formations of multiple unmanned vehicles to traverse between goal points in a dynamic environment has been proposed. The formation is allowed to dynamically change in order to avoid obstacles in the environment while minimizing a cost function aimed at obtaining collision-free and deadlock-free paths. The framework combines global search and local optimization. The global path for the leader of a group is generated via a graph search method. Algorithms are developed to generate smooth trajectories from discrete path sequences. The trajectory planner optimizes the trajectory locally to satisfy the dynamics and input constraints of the individual vehicles while accounting for inter-vehicle collisions and path constraints. A Lyapunov-based controller is designed to keep the vehicles on their planned trajectories. The framework has been extended to multiple unmanned vehicles with trailers. Formation following of a simulated autonomous small grain harvesting system is presented. A powerful set of hardware-in-the-loop software tools for developing and testing behaviors of a group of robots is designed. Illustrative experiments of groups of unmanned vehicles show promise of this approach.