Abstract

In this dissertation, motion coordination strategies are proposed for multiple mobile agents over an environment. It is desired to perform surveillance and coverage of a given area using a Voronoi-based locational optimization framework. Efficient control laws are developed for the coordination of a group of unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) with double-integrator and non-holonomic dynamics. The autonomous vehicles aim to spread out over the environment while more focus is directed towards areas of higher interest. It is assumed that the so-called ``operation costs'' of different agents are not the same. The center multiplicatively-weighted Voronoi configuration is introduced, which is shown to be the optimal configuration for agents. A distributed control strategy is also provided which guarantees the convergence of the agents to this optimal configuration. To improve the cooperation performance and ensure safety in the presence of inter-agent communication delays, a spatial partition is used which takes the information about the delay into consideration to divide the field. The problem is also extended to the case when the sensing effectiveness of every agent varies during the mission, and a novel partition is proposed to address this variation of the problem. To avoid obstacles as well as collision between agents in the underlying coverage control problem, a distributed navigation-function-based controller is developed. The field is partitioned to the Voronoi cells first, and the agents are relocated under the proposed controller such that a pre-specified cost function is minimized while collision and obstacle avoidance is guaranteed. The coverage problem in uncertain environments is also investigated, where a number of search vehicles are deployed to explore the environment. Finally, the effectiveness of all proposed algorithms in this study is demonstrated by simulations and experiments on a real testbed.