Data-driven Unfalsification based Robust Fault Tolerant Control

The paper combined the unfalsified control (UC) which is a data-driven method and Youla parameterization for fault-tolerant control (FTC). The controller reconstruction of the FTC method uses the Youla parameterization for handling with faulty mode. It is a passive and model-free FTC method without know about plant models or faults model. The candidate controller structure is pre-designed by the prior information, and then the FTC is done through tuning Youla parameter without changing the controller structure. The UC-FTC is done by the following three steps: design of candidate controllers, performance monitor, switching and tuning. The method also provides an finite controller set to seek for the best performance by unfalsified control. Finally, the UC-FTC method is applied to a MIMO system.