Synergetic Learning-Based Decentralized Near-Optimal Fault Tolerant Control for Nonlinear Interconnected Systems

This paper proposes a decentralized near-optimal fault tolerant control (FTC) method for nonlinear interconnected systems with actuator faults based on synergetic learning (SL) approach. By incorporating fault signals and control input signals into a loss function and using the SL approach, the system iteratively evolves to eliminate the impact of faults. First, the idea of substituting actual states with desired ones in the coupled system is used to relax the assumption in existing research that the upper bound of interconnected terms must be known, and a neural network observer is constructed to identify unknown interconnected term. Then, considering the different effects of fault signals and control inputs on the system, the FTC problem is transformed into a synergetic adversarial problem between these two signals. Moreover, an improved cost function is designed for the subsystem, and an adaptive critic network with asymptotic stability is established to solve the Hamilton-Jacobi-Isaacs equation, yielding a synergetic approximate solution for the control input and fault assistance signal. Finally, simulation experiment validates the effectiveness of the developed method.