UC San Diego

The use of unmanned vehicles in exploration and surveillance operations has become evermore pervasive in today's world. The development of cooperative motion strategies has been fueled by this increasing demand. However, many dynamical models for these autonomous vehicles remain simple and are not accurate representations of a vehicle where such cooperative motion strategies may be physically implemented. This dissertation will focus on the problem of cooperative deployment of autonomous vehicles subject to various constraints. We will enforce communication or sensing range limitations between agents via range-limited partitions of the coverage environment. Additional novel work includes the investigation of constraints due to remaining power supplies, nonholonomic dynamics, and constraints due to external environmental forces. In addition to these various constraints, we develop a distributed deterministic annealing algorithm to address the separate problem of suboptimal convergence that is exhibited by many of the cooperative deployment problems studied in this work. Theoretical convergence results are obtained in each topic and these are verified in simulation as well as in a robotic test bed developed for this dissertation that we have at UCSD when applicable