Research On Error Suppression Technology Of The Low-cost GPS/INS Seamless Integrated Navigation System

In recent years,the Global Positioning System/Inertial Navigation System(GPS/INS)integrated navigation system has been widely used in unmanned platforms because of its good complementary characteristics.Low-cost integrated navigation technology has become an important development direction of integrated navigation system at present.The GPS/ INS integrated navigation system based on low cost has specific advantages and meets the requirements of the future unmanned vehicle.Meanwhile,combined sensor data analysis technology and the modern data fusion technology,it can realize independent,seamless and high-precision combined navigation system.However,the output noise of the low-cost inertial measurement unit is too larger to obtain valid navigation result.Therefore,researching the error source and error suppression technology of low-cost GPS/INS integrated navigation system is one of the development directions of integrated navigation technology in the future.This thesis studies the low-cost GPS/INS seamless combined navigation error suppression technology,from low-cost GPS/INS seamless integrated navigation system error analysis and modeling,based on fuzzy interval threshold empirical mode decomposition(EMD)gyro de-noising method,based on adaptive complementary filter attitude estimation method and the integrated navigation information fusion method based on the improved radial basis neural network are studied in four aspects.The specific research methods and innovations are:1.For the low-cost inertial device adopted in this thesis,on the one hand,the integrated navigation system model is simplified,and the basic framework of the combined navigation system is determined.On the other hand,the main error sources of the system are analyzed and the error suppression scheme is determined.2.An EMD gyro de-noising method based on fuzzy interval threshold is proposed.The error characteristics of MEMS gyroscope are analyzed.The empirical mode decomposition method is applied to the de-noising of gyroscope output data.Combined with the characteristics of intrinsic mode function obtained by decomposition,the threshold coefficient de-noising is improved,and the fuzzy interval threshold and de-noising method of empirical mode decomposition are proposed.Finally,the simulated signal test and sensor signal test are designed.3.Attitude estimation method based on adaptive complementary filter.The principle of attitude update in low-cost inertial navigation solution is analyzed.The structure of the complementary filter is adjusted by the carrier dynamic information measured by the accelerometer,the accuracy of the system attitude estimation is improved,and the interference of the application environment to the system solution is reduced.4.Inertial error suppression method based on improved radial basis neural network.The adaptive quantum particle swarm optimization algorithm is used to improve the training precision and generalization ability of the radial basis neural network.This network is introduced into the adaptive kalman filter to train the output observation vector.When the satellite signal outages,the divergence of system error is effectively suppressed.Finally,the thesis carries out simulation test and vehicle dynamic test through the self-developed positioning system,which proves the reliability and effectiveness of the proposed method,and provides a reference for the research of low-cost GPS/INS seamless combined navigation error suppression technology.