Research On Static And Dynamic Characteristics And Temperature Compensation Of Six-axis Force Sensor

With the development of industrialization,robot technology has been developed by leaps and bounds.As an important medium for robot to sense spatial full-force information,the application of six-axis force sensor is becoming more and more extensive.With the trend of intelligent and refined robot operation,the demand for high performance six-axis force sensor is becoming more and more intense.This dissertation focused on the most important static,dynamic and temperature characteristics of the six-axis force sensor,and conducted in-depth research from structural principles,static calibration,dynamic modeling and compensation,temperature drift and compensation,etc.The main contents of this dissertation are as follows:1.In view of the problem of static characteristic research of six-axis force sensor,this dissertation built a static calibration system of the sensor,designed two calibration schemes of unidirectional loading and multi-directional mixed loading,and analyzed the coupling characteristic of the sensor through the experimental data.The calibration matrix was obtained by using the least squares and the linear neural network respectively,and the static index of the sensor was further calculated and analyzed.The results showed that the calibration matrix of multi-direction mixed with loading data processed by linear neural network has the best performance.Moreover,maximum linearity error of sensor is 0.96,the maximum repeatability error is 0.66%,the maximum zero drift is 0.12%,all static indicators were consistent with design expectations.2.In view of the problem of dynamic modeling and compensation of six-axis force sensor,this dissertation proposed an improved differential evolution algorithm combined with functional link artificial neural network(IDE-FLANN).On the basis of inheriting the special structure of functional link artificial neural network(FLANN),IDE-FLANN further developed the optimization modeling capability of the differential evolution algorithm(DE).Compared with other optimization methods,IDE-FLANN has the advantages of fast modeling speed,high accuracy and strong robustness.Used the proposed method to study the dynamic compensation of the sensor,taking Mx bridge as an example,the transient time was declined from 512.48 ms to 14.72 ms,and the overshoot amount was declined from 16.8% to 6.33%.3.In view of the problem of temperature drift and compensation of six axis-force sensor,the causes of temperature drift of the elastomer part and the conditioning circuit part of the sensor were analyzed in this dissertation,the thermal zero drift and thermal sensitivity drift experiments of the sensor were designed and completed,which proved that the temperature drift of the sensor mainly comes from the elastomer part.Two methods of piecewise function and extreme learning machines(ELM)were used to build sensor temperature compensation model.The results showed that the compensation model based on ELM was more effective,compared with before compensation,the thermal zero coefficient and the thermal sensitivity coefficient of the sensor were declined by two orders of magnitude.