Neuro-fuzzy Control of a Position-Position Teleoperation System Using FPGA

Abstract

This paper presents an ANFIS (adaptive neurofuzzy inference system) controller for a teleoperation system using FPGA (Field Programmable Gate Array). The proposed controller allows adapting to the dynamic variations of the master and slave models by adjusting the output parameters of the neuro-fuzzy network using a learning algorithm, while taking advantage of the benefits of the FPGA computing power and its high sampling frequency. The ANFIS controllers are developed in MATLAB-Simulink environment and implemented using Simulink's Fixed point tool and HDL Coder. These features provide a fast and accurate control algorithm while optimizing the hardware resources used by the FPGA. The proposed controllers are implemented on a teleoperation system with one degree of freedom. The experimental position tracking results clearly show that the proposed control algorithm guarantees better performance compared to conventional control methods (PID).