Design And Implement Of Beidou/SINS Integrate Dnavigation System Based On ARM Cortex-M7

With the gradual improvement of Beidou satellite navigation system and the development of Beidou/SINS integrated navigation technology,Beidou/SINS integrated navigation system has been applied in aviation,precision agriculture and other fields,but the problem of combined navigation accuracy degraded when Beidou satellite signal out of lock is still one of the current research hotspots.The paper carried out the following research on the miniaturization of the Beidou/SINS integrated navigation system and the problem of sharp decline in navigation accuracy of the navigation system when the Beidou satellite signal is out of lock.Firstly,the basic principle of the strapdown inertial navigation system and the derivation process of the error equation are studied.The existing data integration of Beidou/SINS loose combined navigation based on KF and SVR is studied.The scheme is improved,and the velocity error estimated by the Kalman filter is used as the training target,and the training data is updated in real time.Secondly,the basic principles of Kalman filter algorithm and SVR algorithm are studied.Based on this,the state space model of Kalman filter is derived.The implementation principle of Sequence Minimum Optimization training algorithm in SVR is studied.The hybrid remel frution which is applicable to train data is constructed.The grid cross-optimization algorithm is used to optimize the main parameters m and ? online which affect the performance of SVR prediction.Paraments m and ? combined with the offline C and ? parameters,the SMO algorithm is used to train the speed error SVR prediction model.Thirdly,the usage of the ARM Cortex-M7 core-based microcontroller STM32F767 and ADIS16405 micro-inertial measurement unit and UB280 Beidou satellite navigation receiver are studied.The Beidou/SINS loose integrated navigation hardware system is built and the hardware system is tested.Finally,the integrated navigation system software is designed according to the improved SVR/KF data fusion scheme.The software includes subroutines such as SINS solution,Kalman filter,SVR training data storage,SVR parameter optimization,SVR training and SVR prediction.At the same time,the operation speed and accuracy of the integrated navigation algorithm are guaranteed by using the DSP instruction library of STM32F767.The experimental results of the turntable and sports car show that the improved SVR/KF data fusion scheme can not only solve the problem of sharp drop in the accuracy of the combined navigation when the Beidou satellite signal is out of lock,but also is achievable on the ARM Cortex-M7 microprocessor.The research results of the thesis provide a reference for the optimization of the integrated navigation system algorithm and the selection of the microprocessor.