Model free adaptive fault-tolerant consensus tracking control for multiagent systems

The model free fault-tolerant consensus tracking problem is investigated for multiagent systems subjected to actuator faults. For actuator fault detection, an auxiliary variable based model free fault detection algorithm is introduced to detect the faults, which is capable of detecting the faults occurring on different agents through any agent. The symmetric Laplacian matrix and known system dynamics are not required. When the fault is detected, radial basis function neural network is applied to estimate actuator faults. Next, the fault estimations are utilized to achieve controller reconstruction for the faulty agent. A distributed model free adaptive fault-tolerant consensus control method is developed to guarantee that all agents are able of tracking the expected trajectory. Compared with previous methods, the designed distributed fault-tolerant control method relies merely on agents' input/output data and does not need Laplacian matrix to be symmetric. At last, a simulation is given to validate the effectiveness of developed method.