Hardware Of GPS/DR Integrated Vehicle Navigation System And Piezoelectric Gyro Signal Processing

With the popularization of the automobiles Traffic becomes to be a much moreserious problem, to which GPS navigation is a good solution. But under someconditions the dynamic performances of GPS fall sharply, it may lose it orientatingaction because the navigation satellite is sheltered by buildings. This problem restrictsits development. So a kind of GPS/DR integrated vehicle navigation system withmoderate price and high precision is presented in this thesis, and gains a primaryachievement.Firstly, the thesis describes the structure of the GPS/DR integrated vehiclenavigation system, which integrates the sensors such as GPS, piezoelectric gyro,electronic compass, accelerometer and mile meter. Based on the microprocessorARM9 the modularized design method is realized in each functional hardwaremodules, the principles of which are described in detail.Secondly, the thesis analyses the CMOS chip KMZ52 and relative error sources,and then designs an advanced electronic compass circuit. This circuit adoptsDarlington transistor trigger to ensure the high sensitivity of the inductive components,and adopts MAX392 to make the two output voltage match each other. What's more,the integral negative feedback is realized to compensate the temperature driftaccording to temperature. The interference field calibration and circle runningexperiments prove that the errors are compensated and the azimuth matches with theactual data.Thirdly, the formulas of pitch and roll and the calibration of the horizontal accelerometer arededuced by applying a vertical accelerometer. This accelerometer can output both stance(姿态)parameter and precise acceleration.Finally, a new scheme of dynamic zero correcting for piezoelectric gyro is presented and thediscrete time model of piezoelectric gyro is established, combined with the character ofpiezoelectric gyro and integrated vehicle navigation system, adaptive Kalman filteralgorithm with fault-tolerant ability is proposed by analyzing normal Kalman filterand fault-tolerant Kalman filter, which has obtained satisfied results in the filtering ofactual output data.