Research On Indoor User Moving Trajectory Matching Algorithm Based On Intertial Data

Currently, WLAN positioning based on fingerprint technology is becoming the first option for world-wide indoor positioning service. However, the biggest obstacle that hinders the promotion of WLAN indoor positioning is the labor and time cost of building Radio Map. So how to guarantee the expected positioning accuracy, when improving the speed of building Radio Map to save the labor and time cost, has become a focus research at present.This paper mainly studies how to automatically build Radio Map by indoor user moving trajectory matching algorithm based on inertial data. In order to both reduce the cost of radio map building and ensure the positioning accuracy. Research is focused on how to turn the user relative moving trajectory based on PDR into user absolute moving trajectory, and then combined with the received Signal Strength to complete the establishment of Radio Map, at last complete the online position.For the problem of existing matching algorithms is poor applicability. This paper proposes a rotation matrix automatic generation method, which mainly use Hough transform, Harris corner detection, computational geometry and other technology. It can greatly reduce the manual effort to establish a matrix and simplify the matching process and the complexity when it used in matching algorithm indoor.For the problem of the traditional PDR method could not offer the absolute user moving trajectory to provide the location space of Radio Map. This paper proposes a indoor user moving trajectory matching algorithm based on inertial data to provide the reference point’s position by matching the trajectory generated from the PDR method with all the possible trajectories within the floor plan and could reduce the workload of radio map building by people.The simulation results show that the proposed algorithm could automatically provide the absolute position for the PDR-based trajectory and be robust to the estimation error of PDR method. In order to evaluate position performance of the Radio Map, in operating phase this paper adopts the K Nearest Neighbors algorithm(KNN) for fingerprint matching. Experiment results show that the proposed trajectory matching algorithm in this paper achieves comparable location accuracy to previous approaches based on mobile crowdsourcing even without site survey. It can guarantee the expected positioning accuracy, when improving the speed of building Radio Map to save the labor and time cost.