Study On Indoor Positioning Method Based On PDR And Map Matching

With the rapid development of the digital age,people's demand for location services in indoor scenarios is becoming stronger.However,satellite navigation systems cannot meet user needs in indoor environments because of signal occlusion and interference.There are a variety of indoor positioning technologies have been studied.Among them,pedestrian dead reckoning technology based on mobile smart terminal sensors is favored by researchers because of its strong autonomy and continuity.However,PDR relies on the implementation of the built-in sensor of the smart phone,and is limited by factors such as size,cost,power consumption,etc.,so its measurement accuracy is limited,which causes positioning errors to accumulate over time.In order to solve this problem,this paper fuses the map information and motion behavior with positioning data to correct the position estimation results and weaken the accumulated error of PDR positioning.Then the positioning result is calibrated to the traffic path to achieve the purpose of optimizing the position output and improving the user experience.The main contents of this article are as follows:(1)Pedestrian dead reckoning is realized based on smart phone MEMS.The initial location is determined by Bluetooth location fingerprint matching technology.Based on acceleration data,gait detection algorithms such as peak detection,zero-crossing detection and correlation coefficient detection are used to achieve accurate step counting.Step length estimation models such as constant model,linear model,nonlinear model,empirical model and zero-speed correction model are studied,and some improvements are made to the step length estimation model on this basis.Euler angle method and quaternion method are used to estimate pedestrian heading based on gyroscope and magnetometer data.(2)An indoor navigation path map for map matching is constructed.Indoor space is divided into two types: vector domain space and non-vector domain space.An indoor navigation path map oriented to map matching is constructed based on the above two spatial types.The construction method of navigation path graph in indoor scenes with unknown spatial layout have been studied.(3)Pedestrian dead reckoning is realized based on indoor virtual landmarks.The auxiliary effect of indoor virtual landmarks based on wireless network environment and pedestrian behavior characteristics on pedestrian dead reckoning is studied.The correction of PDR positioning results by Bluetooth beacons and corner landmarks is realized,which has a positive effect on reducing accumulated errors.(4)A path matching assisted PDR location algorithm based on multiple weights is proposed.The limitations of traditional projection-based path matching algorithms are analyzed,and multiple influencing factors such as distance,direction,connectivity and human flow are integrated into the optimal matching path evaluation process on the basis of it.A path matching algorithm based on multiple weights and a weight coefficient model considering different scenarios are proposed.Experiments show that the path matching algorithm based on multiple weighting improves the average positioning accuracy by 33.7% compared with the traditional PDR algorithm,and the maximum positioning error is reduced by 43.9%.