Research Of Train Integrated Positioning System Based On Data Fusion

Train location information is one of the important parameters in train control system, train positioning method based on multi-sensor fusion technology is an effective way to improve the accuracy of train positioning. Global positioning system and inertial navigation system can form complementary advantages, therefore, the method combines GPS and iner-tial system together, and appling it to train positioning system, the method can avoids the promblem of using single positioning technology, and effectively improve the accuracy and reliability of positioning informations. Traditional sensor fusion technology estime the loca-tion information by using Kalman filter, however, this estimation algorithm needs a premise which the accurate system model and the statistical properties of the exact external interfer-ence signal must be known, because of these limitations, in the actual train operation process, the accuracy of the location informations which estimate by Kalman filter algorithm will be affected, in severe cases, the filter may occur divergence problem.In order to overcome the above issues, this thesis started on the way of train integrated positioning, selected the appropriate sensors, designed integrated positioning model, and focused on the research of sensor information fusion method which applied to train integrated positioning. On the basis of analysising the performances and features of the current navigation system, then choose the GPS and SINS as the subsystems of the integrated posi-tioning system, thus avoid the defects of using single positioning technology and achieve the complementary advantages. When establishing the subject of GPS/SINS integrated position-ing model, the thesis simplifies the SINS system according to the actual working conditions under the railways, and the integrated model uses position and velocity combined mode. The thesis analysis the data fusion algorithms which are being used in the integrated navigation system, in view of the train actual operation, the system model and external disturbance char-acteristics can not be informed accurately, therefor the thesis discussed the method of data fusion which adopted the H∞robust filtering algorithm, and comparing with the traditional Kalman filtering algorithm, the simulation results verifies that the H∞robust filtering algorithm relative to the Kalman algorithm can obtain better filtering effect.On the basis of SINS/GPS integrated positioning, in order to further improve the train positioning accuracy and amend the positioning informations. According to the characteristics of railway transport, the thesis designs a map matching algorithm based on GPS to accelerate the search speed. In the thesis, road network block and the principle of minimum difference are introduced in the map matching algorithm, the experiments show that the algorithm is feasible.