The Research And Realization Of Accelerometer Automatic Calibration System

Accelerometer is one of the important branches of sensor, which plays animportant role in all walks of life.With the development of AutomotiveElectronics,Consumer Electronics,and etc,demand of Accelerometer is growing with ahigher requirement to accuracy.The Accelerometer must be calibrated before leavingthe factory.Accelerometer production efficiency will be affected without a highefficient,high precision calibration-instrument.This paper designs a set ofAccelerometer automatic calibration system,and the Accelerometer calibrationalgorithm has been studied.The system is based on the principle of centripetalacceleration.standard acceleration is generated by centrifuge which is used tocalibrate the Accelerometer.Data Acquisition Module collects the output signal of theAccelerometer and sent it to the Upper Computer Software.And finally we calibratethe signal of Accelerometer by using calibration algorithms of the Upper ComputerSoftware.To begin with,based on the analysis of principle of Accelerometer and calibrationmethod,we design the overall structure of the system.The system are mainlycomposed of centrifugal machine platform,acceleration transducer module,dataacquisition module,data staging module.Machine structure and communication of thesystem are explained in this paper.Both software and hardware designs are introducedin detail.Meanwhile,hardware design mainly includes microcontroller circuit,RFcommunication circuit,AD converter circuit,Accelerometer channel selectioncircuit,power circuit and etc,while software design mainly contains main programmes,communication program of lower computer and serial port communicationsubprogram,data acquisition subprogram,centrifuge control subprogram,calibrationsubprogram of upper computer software.Furthermore, common calibration algorithms of Accelerometer are further studied andtheir merits and demerits are analyzed. These algorithms include least square method,Back-Propagation(BP)neural network and genetic neural network. We find the problem ofslow convergence,long training time existing in the later period of genetic neuralnetwork.Therefore,a new improved calibration algorithm of GA-PSO optimized BP neuralnetwork is presented. Realizing advantage complementary by Combining particle swarmoptimization algorithm,genetic algorithm and Back-Propagation neural network, the proposed algorithm improves the computational efficiency and precision of calibration algorithmduring calibrating Accelerometer.Finally, the calibration experiment of SCA60C accelerometer using the automaticcalibration system are carried out. The calibration result of SCA60C accelerometerwith different algorithms were detailed analysised and compared. The experimentalresults shows that the improved calibration algorithm has better precision andconvergence speed compared with genetic neural network calibration algorithm,andthe effectiveness and feasibility of Accelerometer automatic calibration system isverified.