Research On Fusion Positioning Strategy For Land Vehicles In Satellite Signal--blocked Traffic Environments

The integration of Global Positioning System(GPS)and Inertial Navigation System(INS)based on Micro-Electro-Mechanical System(MEMS)technology provides a low cost solution for accurate and reliable land vehicle positioning.However,the accuracy of MEMS INS/GPS degrades sharply during GPS outages.To deal with this problem,fusion positioning strategy for land vehicles in satellite signal-blocked traffic environment is in-depth studied in this paper.This paper focus on two key technologies,including random error suppression of Micro-Electro Mechanical Systems(MEMS)Inertial Measurement Unit(IMU)and information fusion algorithm for integrated navigation,and carry out detailed research for some key issues.The main research work and fruits are summarized as follows:(1)To eliminate the high-frequency part of MEMS IMU random errors,an inertial data preprocessing algorithm based on improved wavelet filter(IWF)is designed in this paper.Aiming at the drawbacks related to conventional wavelet filter,this algorithm achieves the optimal level of wavelet decomposition based on spectrum analysis and experimental evaluation,and designs an enhanced threshold function with an adjustable coefficient.Experimental results demonstrate that the IWF is capability of overcoming the shortcomings of traditional threshold function to some extend,and effectively removing the high-frequency part of random errors in MEMS inertial sensor measurements.(2)With the purpose of suppressing slowly varying part of MEMS IMU random errors,this paper develops an adaptive data preprocessing algorithm based on Empirical Mode Decomposition(EMD)interval threshold filter,which is designed by integrating the advantages of EMD and fractional Gaussion noise(fGn).This paper adopts fGn to model the inertial sensor random error,and selects the EMD filter threshold according to the variance relation among each Intrinsic Mode Function(IMF).In addition,a new threshold shrinkage scheme based on IMF interval is developed to eliminate the discontinuity of filtered signals.Experimental results show that the algorithm can effectively and efficiently suppress not only partial slowly varying random errors,but also the high-frequency part of random errors.Compared with IWF,this method accomplishes a further improvement in the accuracy of filtered signal.(3)In order to improve the accuracy and raliability of vehicle integrated positioning system in satellite signal-blocked traffic environments,a hybrid strategy based on Auto-Regressive(AR)model aided Kalman Filter(KF)is proposed.First,this method introduce inertial data preprocess step to the system structure in order to provide more accurate data for subsequent information fusion.Then,a new INS error series modeling and prediction structure is developed based on the modification of traditional INS error modeling.On the basis of these improvements,an autoregressive model(AR model)-based forward estimator(ARFE)is designed and augmented with KF,known as ARFE/KF hybrid strategy,to predict and compensate INS position errors.Real road test results demonstrate that this method can efficiently compensate INS positioning error and significantly enhance the accuracy of vehicle positioning during GPS outages.In addition,this method owns outstanding performance in generalization ability and real-time capability.(4)In order to further improve the performance of vehicle integrated navigation system in relatively long GPS outages,this paper proposes an improved hybrid strategy,in which the KF is aided by a Least-Squares Support Vector Machine(LSSVM)-based nonlinear autoregressive with exogenous input(NARX)model(LSSVM-NARX).Taking account of historical development trend of INS error and the vehicle running states,this method firstly designs a system structure with memory and inner feedback to model INS positioning errors.Then,a LSSVM-NARX/KF hybrid strategy is developed to model INS positioning errors,and compensate INS positioning errors during GPS outages.Experimental results show that this method has good adaptability to various driving conditions,and efficiently suppresses the accumulation of INS positioning errors,thus providing accurate and reliable vehicle positioning information in relatively long GPS outages.