Interval Optimal Controller Design for Uncertain Systems Based on Interval Neural Network

This paper presents a novel method for designing interval optimal PID controller for uncertain systems, which utilizes the key technologies of interval BP neural network (IBPNN) and interval particle swarm optimization algorithm (IPSO). The IBPNN is employed to model the uncertain plant as an interval system, and the output of the interval PID is applied to this interval network. By using the IPSO, the parameters of the interval PID are optimized, resulting in the determination of the parameter domain. This method effectively tackles the challenges associated with designing interval controllers based on Kharitonov's theorem, such as the dependence on interval plant models, intricate calculations, and the inability to handle nonlinear plants. Moreover, it circumvents the issue of interval expansion caused by interval operations. By ensuring the stability of the interval system, the proposed method can also optimize the system's performance. The simulation results validate the effectiveness and generalization of the proposed method.