Research On Map Matching Technology Based On Route Features

Accurate and real-time position is not only the premise of vehicle transportation safety and efficiency,but also the basis of intelligent transportation.In order to improve the accuracy of vehicle position and adaptability to complex scenes,this paper designs a map construction and map matching method based on the electronic map combined with the route features,realizes the map aided vehicle position method.On this basis,the simulation experiment and vehicle test scheme are designed to verify the model,and the correction effect of the model on the position error is analyzed.This paper mainly carried out the following aspects of research:(1)Study the theoretical basis of map matching and vehicle positioning,analyze the plane line characteristics,and design parameter equation and map plane model according to the parameter model of straight line,transition curve and circular curve,which are the basic model of map construction and map matching research.Then,the research is carried out from two aspects of map construction and map matching,in which map construction is realized by line segmentation and trajectory fitting,and map matching is carried out from the perspective of multi-source information fusion.(2)According to the autonomy and attitude integrity of INS,a two-layer threshold identification method of planar linear based on INS is proposed.According to the characteristics of heading angular velocity in different alignment in driving state,a double threshold extraction algorithm is designed to divide straight line,transition curve and circular curve,and the recognition results of three types of alignment are obtained.The domain grid clustering is used to correct the wrong section division and line type misjudgment caused by the outliers.Through the above two-layer threshold extraction algorithm to achieve the research on line segmentation model based on INS attitude information.(3)On the basis of route segmentation,the trajectory fitting method is designed.Due to the interference of test conditions and route environment,the coupling density coefficient is introduced to improve the adaptive weight of the standard LS,and the fitting coefficients of the improved LS fitting algorithm are derived according to the route features,including the fitting coefficients of straight line,circular curve and transition curve.The fitting trajectory of each road section is obtained,and the topological map is constructed according to the map model by extracting the intersection points.(4)This paper proposes a map matching model based on route features,and carries out research from three aspects: Initial section matching,initial position and direction matching,section matching and intra section tracking matching.In the initial matching process,a matching method based on system stability discrimination is proposed;based on the section matching,the attitude constraint model is designed to improve the GNSS/INS algorithm,and dead reckoning in map by MM/DR,combine the GNSS/INS/MM by improved FKF integrated algorithm.Thus,in order to improve the accuracy of vehicle position,a track matching method based on route features is proposed to correct GNSS/INS integrated navigation system.(5)After completing the initial matching,the tracking matching stage is started,and the tracking matching model based on route features is designed.The model consists of three parts: Firstly,the GNSS / INS algorithm is improved by designing the attitude constraint model under the line feature;Secondly,the MM/DR position information is formed by one-dimensional mileage calculation,Finally,the improved FKF integrated navigation algorithm is used for multi-source information fusion,the sub filter is designed according to the attitude constraint model,and the main filter of adaptive information distribution is designed.The information distribution coefficient of each filter is adjusted in real time according to the driving environment,tracking matching model is used to improve the accuracy of vehicle position estimation and adaptability to driving environment.(6)Finally,the simulation experiment and vehicle experimental scheme are designed to verify the plane alignment recognition based on the double threshold model,the improved adaptive LS fitting algorithm and the map matching model based on the route features.The experimental results show that,the accuracy of the line segment and road segment recognition method based on the double threshold recognition model is about 97.1%,the fitting error of the improved adaptive LS section fitting method is reduced by 65.21%,the map matching model based on route features can reduce the error of position and speed estimation by more than 50%.