Trajectory planning and stabilization for formations acting in dynamic environments

Abstract

A formation driving mechanism suited for utilization of multi-robot teams in highly dynamic environments is proposed in this paper. The presented approach enables to integrate a prediction of behaviour of moving objects in robots' workspace into a formation stabilization and navigation framework. It will be shown that such an inclusion of a model of the surrounding environment directly into the formation control mechanisms facilitates avoidance manoeuvres in a case of fast dynamic objects approaching in a collision course. Besides, the proposed model predictive control based approach enables to stabilize robots in a compact formation and it provides a failure tolerance mechanism with an inter collision avoidance. The abilities of the algorithm are verified via numerous simulations and hardware experiments with the main focus on evaluation of performance of the algorithm with different sensing capabilities of the robotic system. © 2013 Springer-Verlag.