Research On Indoor Fusion Positioning Algorithm Based On Intelligent Terminal

The indoor positioning system with single data source is difficult to meet the demand for location-based service.Therefore,efficient fusion indoor positioning technology has become a very active research topic.However,the positioning accuracy of the fusion positioning system is still limited by the accuracy of each sub-positioning system.Improving the positioning accuracy of each independent positioning system and then using the fusion positioning technology can obtain better positioning result.This thesis improves the Bluetooth-based positioning algorithm and Pedestrain Dead Reckoning(PDR)positioning algorithm respectively,and realizes the indoor fusion positioning algorithm based on intelligent terminal.The specific work is summarized as follows:1.A PDR algorithm based on motion mode recognition is proposed.Three issues of course estimation,gait detection and stride length estimation in the PDR algorithm based on intelligent terminal are studied.Since the user's step size is related to its motion mode,The thesis proposes a PDR algorithm,which uses Dynamic Time Warping(DTW)to recognize users' motion mode and estimates the stride length according to the motion mode.This algorithm can avoid feature selection and complex calculations.The experiment result verifies that the proposed algorithm can further improve the system performance.2.A Received Signal Strength Indication(RSSI)real-time correction algorithm is proposed.The existing Bluetooth-based indoor positioning algorithm is studied to analyze the causes of errors.Based on the error analysis of the RSSI collected by the intelligent terminal,a RSSI real-time correction algorithm used Bluetooth gateway is proposed.The algorithm can obtain smoother and stable RSSI data under the condition of satisfying lower delay.The experiment shows that the proposed algorithm can reduce the positioning error caused by RSSI fluctuation.3.A propagation path loss model based on PSO-BPNN(Particle Swarm Optimization-Back Propagation Neural Network)is proposed.The PSO-BPNN is used to train the RSSI distance model to accurately estimate the distance between the blind node and the Bluetooth beacon.4.Research on multi-sensor fusion positioning algorithm since indoor positioning system with single data source has low positioning accuracy.The particle filter algorithm is used to fuse the result of the PDR positioning algorithm based on motion pattern recognition and the Bluetooth positioning algorithm based on improved RSSI path loss model.Finally the indoor fusion positioning algorithm based on intelligent terminal is realized.The experiment shows that the proposed algorithm can further improve the system performance.