The Calculation Of Motion State And Real-time Simulation Based On MEMS

With the development of MEMS technology, the micro measurement system have beenapplied more and more frequently. It has attached the attention of many researchers especiallyin motion measurement field, such as unmanned vehicle navigation, robot control, etc. Thetechnology based on theory of inertial navigation, but it has its own characteristics andappropriate programs are needed for different applications. In this paper, a micro motionmeasurement system including a MEMS gyroscope, a MEMS accelerator and a magnetometerwas designed, the resolving scheme and the simulation methods had been discussed in detail.Firstly, this paper introduced the development of related technologies in China andabroad, including the development of MEMS inertial component, the research about attitudecalculating algorithm and some outstanding micro motion measurement system instances.Secondly, this paper discussed various kinds of sensor data processing methods. Based on thecharacteristics of MEMS gyroscope, MEMS accelerometer and magnetometer, theappropriate error compensation and filtering schemes were achieved. The data processingmethods for MEMS gyroscope using the Allan variance analysis, AR model, kalman filteringtheory was the emphasis. And the medfilter and a kind of threshold method based on DFTwere imported for data filtering of accelerometer and magnetometer. Thirdly, the attitudealgorithm was discussed in detail. The quaternion algorithm, the theory ofaccelerometer/magnetometer integrated attitude measure system, the data fusion methodsbased on the kalman filter and the complementary filter are introduced. And then, the speedcalculating algorithm was achieved with HDR introduced for the reduction of bias of linearacceleration.In order to verify the effectiveness of the various solutions designed in this paper, a micromotion measurement system hardware based on ARM was realized. And a software whichcan draw the curves and operate3-D graphic simulation on-line was successfully developedin visual C++6.0. At last, the best solving scheme was achieved based on the theory andexperiments.The sensors used in this paper are very cheap, so the measurement results are not veryreliable, this shortage is made up compensated by the optimization of the algorithm. So it’svery useful for many situations.