Date of Award
Open Access Dissertation
Since the consensus problem is getting popular in recent researches and the applications based on it are expanding in current years! the consensus problem under the condition that constraints exist on agents’ state in continuous-time in a multi-agent system has been studied here. This work combines the projection operator method in a general consensus algorithm to avoid the violation of state on its individual constraint set. When the projection operator works in the general consensus algorithm, the constraint region is divided into the projection-operator working region and the non-projection region. The boundary between these two regions is a circular line. Once the state of the correspond agent cross this line, the projection operator starts to work when the state is going out from the non-projection region and moves towards to its constraint boundary. The boundary of the starting circle for projection opera- tor has the same shape with the constraint set boundary. The difference is it has a smaller area with a certain ratio. With a one dimensional projection-based consensus algorithm, we analyzed the projection-based consensus algorithm on the expected-to- arrival (ETA) requirement which is always the problem at the simultaneous arrival task on multiple unmanned aerial vehicles (UAVs) system. The stability and con- vergence rate of the proposed algorithm has been analyzed. Based on the stability and convergence rate, the conditions to guarantee the feasibility of the simultaneous arrival task has also been presented. Then the analysis for projection-based consensus algorithm has been extended in a high dimensions, in this case, the agents can only exchange their state information in continuous time in a connected and undirected communication network topology. The result is proven that since the intersection set of every state constraint is non-empty, the global consensus will eventually be achieved. At last, the projection-based algorithm has been applied on a UAV module, with the UAV dynamics, the algorithm makes a multi-UAV system with different ability in flying reach the global spatial consensus achievement for simultaneous arrival task.
Zhou, Z.(2019). Constrained Consensus in Continuous-Time Multi-Agent Systems. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/5296