In this thesis, the problem of distributed fault tolerant control for multi-agent systems with complex-weights directed communication topology is studied. First, a framework is proposed to associate optimal topology of multi-agent systems to a constrained optimization problem with complex Laplacian matrix, which is independent of the agent dynamics. Second, a bank of observers is constructed to compute residual vectors with an aim of allowing the fault detection and isolation of actuator fault to occur on any agent of the multi-agent systems. Third, a distributed finite-time convergent observer is proposed for each agent to estimate the state and fault in a finite time. Then, using the fault information obtained online, an adaptive fault tolerant control protocol is proposed to compensate for the failure effects and to enable all the agents to achieve the control goal. Also, we show that the closed-loop system can be guaranteed to be asymptotically stable in the presence of faults. Finally, some numerical example results are given to demonstrate the effectiveness of the proposed schemes. Key Words Multi-agent system, Complex Laplacian, Fault tolerant control.
