Formation control of agents is one of the most active research areas in the field of multi-agent control systems which have many applications in various fields of industry, military, and traortation. Importance of operational performance of components, and constitutive elements of these systems in appropriate operation of the whole system brings attention to the issues of detection, isolation, and fault tolerance in multi-agent systems. In the formation of some systems, occurrence of a small fault in one of the agent leads to releasing fault affect in other subsystems and losing desirable performance. So it is of particular importance to consider fault tolerant control of agents in maintaining formation control and also tracking desirable path by agents in order to increasing reliability, safety, and performance. In this thesis, stability of adaptive fault tolerant controllers in formation control of multi-agent systems has been analyzed when its parameters are updated. In some cases, it is not practical to assume a known set of possible faults or upper bounded of faults in the systems, because there is not much information available about fault and its magnitude. Furthermore, systems being made by several agents have large-scale and limited information about neighbors. Thus, in order to solve the mentioned problems it seems essential to propose an adaptive fault tolerant controller for additive fault with unknown bounded. The controller is completely distributed without need to system globally information. At first, with assumption of time-invariant formation, the maintaining formation problem of agents with undirected topology and desirable path tracking problem with directed topology by leader follower approach has been discussed in the presence of additive fault. Considered Dynamic in each cases is linear double integrator. Then with assumption of time-varying formation, the maintaining formation problem of linear high-order dynamic with undirected topology has been investigated. In each of the considered issues, the appropriate adaptive rules have been proposed for updating the fault tolerant controller parameters. The effect of adaptive rules on Asymptotic convergence of absolute and relative position error of agents to zero in the maintaining formation and so uniformly ultimately bounded of these errors in the tracking has been guaranteed by using appropriate Lyapunov function. Key Words :multi-agent systems, fault tolerant, adaptive controller, distributed controller.
