Signal Processing Based On The AD Series Of Gyroscope And Accelerometer And Its Application Research

MEMS inertial sensors has a widely application value in the angle measuringsystem with the progress of science and technology. The precision of inertial systemis mainly depened on the resolution of the gyroscope and accelerometer. A lot ofresearch on precision of inertial system had improved the accuracy by improving thehardware structure and processing technology, although we enhance the precision ofinertial sensors to a certain degree extent, but it also give rise to difficulty of thetechnical and expensive of the cost. So we need to seek a method of software toreplace inertial sensor signal processing system, in order to reduce the processrequirement and cost. Using software optimization algorithm to improve the precisionof the signal is an important research direction of the inertial sensor.Based on the micro mechanical AD series accelerometer ADXL203andgyroscope ADXRS622are used as the research object.This thesis probes thecharacteristics of the micro inertial component and the signal processing method fromthe theoretical analysis, circuit design, algorithm simulation, test, etc. The focus of thethesis researche from two aspects, it includes the design of the micro inertial sensorsmeasuring system and software filtering algorithm. The experiments show thatwavelet threshold denoising algorithm is superior to kalman algorithm of the digitalfilter. It has a good effect in practical application.The thesis main work includes:Firstly, the thesis is based on service behavior and characteristics of constructionof ADXL and ADXRS series of MEMS inertial sensor, we analysis performancecharacteristics which is ADXL series of accelerometer and ADXRS series ofgyroscope, complete specific projecting and physical implementation with MEMSinertial sensor between data acquisition with signal conditioning circuit, and betweenthe processor with A/D conversion interface. Meanwhile, we design the power supplymodule. We use the experiments verify the feasibility of the scheme design.Secondly, we design the gyroscope detection platform and study theaccelerometer ADXL203and gyroscope ADXRS622sensor output data informationprocessing method. TMS320F28335is used as the system processor. We write related program. The thsesis introduces and analyzes the wavelet transform thresholdde-noising algorithm and kalman filtering algorithm. Kalman filtering algorithm isbased on CCS4.2real-time filtering software platform. We simultaneously write the Clanguage of program of kalman filtering algorithm. The experimental results showthat the algorithm of kalman filtering effect has the very big space to improve.Thewavelet analysis algorithm is based on the MEMS gyroscope drift analysis. Wesimulate experiment in MATLAB software and select the optimal wavelet function,decomposition layers and threshold value. In the meantime, we write the wavelettransform threshold denoising program of matlab. Through Many experimental resultsshow that the sym4wavelet function and6layer decomposition and heuristic hardthreshold de-noising effection is best.Finally, the design of the AD series of MEMS inertial sensors measuring systemis applied to two rounds of the balance of the car. We compare with the advantagesand disadvantages of the kalman-filter and the wavelet transform thresholdde-noising.We choose the wavelet transform threshold algorithm for current-timefiltering. The car body control experiments show that wavelet transform thresholdfiltering has good filtering effect for AD series of the sensors. It can effectively inhibitdrift with gyroscope. At the same time, it shows that the real-time algorithm can meetthe needs of controlling with the two-wheel self-balancing, the car body balance effectis better.