Adaptive Neuro-fuzzy inference system based estimation of EAMA elevation joint error compensation

Highlights: •Adaptive Neuro-fuzzy Inference System for estimating the compensation error from an unformulated cluster of data. •The root mean squared error is significant improved, which satisfy the accuracy requirement up to 0.02 °. •Easily incorporate the non-mathematical model sensors calibration and advance control algorithms. -- Abstract: EAMA (EAST Articulated Maintenance Arm) is an articulated serial robot arm working in experimental advanced superconductor tokamak for the inspection and maintenance. This paper implements algorithms to calibrate the synchronize deflection and estimate its signal for the robot control. The retrieval has two distinct tasks, a shaft rotation direction signal processing and a discrete data classification, meanwhile neuro network and expert system are applied for completing these separate tasks respectively. In this paper the use of Adaptive Neuro-fuzzy Inference System for estimating the compensation error from an unformulated cluster of data that has unneglectable nonlinearity is presented. The simulation result shows that the root mean squared error is significant improved, the final results satisfy the accuracy.