Neural iterative learning identifier‐based iterative learning controller for time‐varying nonlinear systems

Abstract

AbstractIn this paper, two iterative learning control (ILC) approaches based on a neural iterative learning identifier (NILI) are proposed for a class of unknown time‐varying nonlinear systems. Since both control and identification processes are done at the common iteration domain, so both tracking performance and identification performance are improved at each iteration simultaneously. A class of neural networks is proposed as iterative learning identifier. The proposed NILI is time varying because the system is time varying. An auxiliary error function is used in new iterative updating laws of the proposed NILI. First, P‐type ILC with iteration‐varying learning gain based on the NILI is proposed and applied to the system. Then, a new ILC based on the NILI is proposed and applied to the system. In comparison with other ILC algorithms proposed for similar systems, the proposed algorithm in this paper has more accurate performance and takes less amount of time and memory. Also, the new proposed NILI‐based ILC can be successfully applied to the system with iteration‐varying initial condition. Theoretical analysis of the trajectory convergence both tracking and identification errors is fully done. Finally, the effectiveness of the proposed algorithms is verified through some simulation case studies.