Fuzzy Self-tuning Tracking Differentiator for Motion Measurement Sensors and Application in Wide-Bandwidth High-accuracy Servo Control.

Yang Gao,Dapeng Tian,Wang Wang

doi:10.3390/s20030948

Abstract

Sensor differential signals are widely used in many systems. The tracking differentiator (TD) is an effective method to obtain signal differentials. Differential calculation is noise-sensitive. There is the characteristics of low-pass filter (LPF) in the TD to suppress the noise, but phase lag is introduced. For LPF, fixed filtering parameters cannot achieve both noise suppression and phase compensation lag compensation. We propose a fuzzy self-tuning tracking differentiator (FSTD) capable of adaptively adjusting parameters, which uses the frequency information of the signal to achieve a trade-off between the phase lag and noise suppression capabilities. Based on the frequency information, the parameters of TD are self-tuning by a fuzzy method, which makes self-tuning designs more flexible. Simulations and experiments using motion measurement sensors show that the proposed method has good filtering performance for low-frequency signals and improves tracking ability for high-frequency signals compared to fixed-parameter differentiator.

Highlights

IntroductionSignal differential estimation is essential for wide-bandwidth high-accuracy servo controls [1]
Signal differential estimation is essential for wide-bandwidth high-accuracy servo controls [1].High-precision measurement of position, velocity and acceleration signals is a necessary feedback information for various control strategies, including classical PID control [2], sliding mode control [3], and so on
The basic idea of adaptively adjusting the g value is obtained: Smaller g values are preferable for low-frequency signals, whereas larger g values are preferable for high-frequency signals

Summary

Introduction

Signal differential estimation is essential for wide-bandwidth high-accuracy servo controls [1]. High-precision measurement of position, velocity and acceleration signals is a necessary feedback information for various control strategies, including classical PID control [2], sliding mode control [3], and so on. Due to the limitation of measurement mechanism, some motion measurement sensors, such as velocity sensors and acceleration sensors, have certain measurement noise which limits the accuracy of the measurement. Additional motion measurement sensors are always limited by system installation space and cost [4,5]. Differential signal is extracted from the given input to realize the design of a high order controller. Classical difference methods obtain the differential signal with the small sampling time. The signal noise in the original signal will be greatly amplified

Methods

Discussion

Conclusion