TSK-Type Self-Organizing Recurrent-Neural-Fuzzy Control of Linear Microstepping Motor Drives

In this paper, a Takagi-Sugeno-Kang-type self-organizing recurrent-neural-fuzzy network (T-SORNFN) is proposed for the trajectory tracking control of linear microstepping motor (LMSM) drives. Without a priori knowledge, the T-SORNFN is constructed to model the inverse dynamics of a LMSM drive by a set of recurrent fuzzy rules built online through concurrent structure and parameter learning. The fuzzy rules in the T-SORNFN can be either generated or eliminated to obtain a suitable-sized network structure, and a recursive recurrent learning laws of network parameters are derived based on the supervised gradient-descent method to achieve fast-learning converge. Based on the Lyapunov stability approach, the convergence of the T-SORNFN is guaranteed by choosing varied learning rates. Furthermore, an inverse-control architecture that incorporates T-SORNFN and a proportional-derivative controller is used to control the LMSM drive in a changing environment. A recursive least-squares (RLS) algorithm is utilized for online fine-tuning the consequent parameters in T-SORNFN to obtain a more precision model. Simulated and experimental results of a LMSM drive are provided to verify the effectiveness of the proposed T-SORNFN control system, and its superiority is validated in comparison with NFN and RNFN control systems.