Intelligent Direct Inverse Control for Nonlinear Systems

In this article, a modified self-recurrent wavelet neural network (MSRWNN) structure is used as a feedforward controller in the direct inverse control (DIC) method. Nonlinear dynamical systems are controlled with the help of this intelligent control strategy. The particle swam optimization (PSO) is used as an efficient optimization tool to determine the ideal MSRWNN parameter settings. A nonlinear dynamical system is considered to demonstrate the efficacy of the suggested control strategy. In addition, the ability of the MSRWNN to effectively regulate the nonlinear system under consideration is specifically assessed in terms of control accuracy and resilience to external disturbances via the execution of many assessment tests. The outcomes of each of these tests have shown the control scheme's effectiveness, with significant improvements in performance metrics. Specifically, the proposed method achieved a 25% reduction in Integral Squared Error (ISE) compared to traditional neural network controllers, and improved disturbance rejection capability by 30%. Furthermore, from a comparative study, the MSRWNN has demonstrated superior control accuracy and performance reliability over other related controllers. Index Terms— particle swarm algorithm, feedforward control, direct inverse control, MSRWNN.