Multi-Source Fusion Indoor Positioning Method For Pedestrians And Vehicles By Smart Phone

With the development of economy,the car has become a common tool for people to walk,and the number of cars has been increasing.The construction of underground parking lot relieves the problem of parking due to the limited urban space,but in the underground parking lot,there are many difficulties in looking for the car or finding the parking space,especially in large complex underground parking lots and less vacant parking spaces.Therefore,the underground parking lot urgently need indoor navigation and positioning technology support.At present,common indoor positioning methods: infrared,ultra wideband and ultrasonic,although the positioning accuracy can reach the level of sub meter,but the cost is high and the coverage is small.Radio frequency identification,ZigBee,Wi-Fi and multi-source fusion positioning accuracy reach the meter level,however,the installation and maintenance are complex and vulnerable to environmental factors,resulting in the current indoor positioning technology coverage is small,and it is difficult to popularize in practical applications.From purpose of reducing the location cost and realizing the convenience to spread positioning algorithm,this thesis studies the indoor positioning algorithm based on Bluetooth,mobile sensor,On-Board Diagnostic and map matching.The experimental results show that under 95% confidence,the positioning error is within1.5 meters.,which meets the needs of indoor people to find his car and his car to find a parking space.The indoor location based on Bluetooth beacon achieves the location accuracy of the room level by distance intersection method.After analyzing the factors that affect the positioning accuracy,combining with indoor road information,projection correction can improve the positioning accuracy of Bluetooth.The experimental results show that the median error of the vertical road is ±3.39 m,and the correction is ±0.17 m.Fusion of Bluetooth beacon,smart phone sensors and indoor map for indoor location.Based on the periodic variation characteristics of heading angle in walking process,a new step number detection method is proposed,and the reliability of the method is verified.In order to realize the dead reckoning when pedestrians swing their arms,the state of the smart phone is judged by three axis acceleration,and the pedestrian heading angle is determined by the relationship between the attitude angles.Integrating indoor map information reduces cumulative error of pedestrian dead reckoning,and the positioning error after fusion decreases from the largest 9m to 3m,with a positioning error of about 95% within 1.5m.Finally,experiments show that pedestrians walk at a constant speed with a horizontal mart phone,only a gyroscope can be used to calculate the pedestrian level.Combined with indoor road information,the positioning error after projection correction can be controlled within 96% of 1.5m.In the multi-source fusion vehicle location,the heading angle of the smart phone and the speed obtained by OBD can calculate the vehicle dead reckoning.The closure error is(-1.263,0.482).In order to improve positioning accuracy and fuse road information,the maximum distance of vehicle departure from reference route after fusion is changed from 9.5m to 3.8m,and the distance from the reference line less than 3m is changed from about 84% to about 98%.