Abstract

In this thesis, different distributed consensus control strategies are introduced for a multi-agent network with a leader-follower structure. The proposed strategies are based on the nearest neighbor rule, and are shown to reach consensus faster than conventional methods. Matrix equations are given to obtain equilibrium state of the network based on which the average-based control input is defined accordingly. Two network control rules are subsequently developed, where in one of them the control input is only applied to the leader, and in the other one it is applied to the leader and its neighbors. The results are then extended to the case of a time-varying network with switching topology and a relatively large number of agents. The convergence performance under the proposed strategies in the case of a time-invariant network with fixed topology is evaluated based on the location of the dominant eigenvalue of the closed-loop system. For the case of a time-varying network with switching topology, on the other hand, the state transition matrix of the system is investigated to analyze the stability of the proposed strategies. Finally, the input saturation in agents' dynamics is considered and the stability of the network under the proposed methods in the presence of saturation is studied.