Pedestrian Indoor Positioning Algorithm Suitable for Unknown Environments

IMU plays an important role in pedestrian positioning due to its small size and light weight, but it also has the problem of accumulated errors. This article improves the traditional fixed threshold zero speed detection algorithm and proposes a valley based zero speed correction algorithm that can dynamically adjust with changes in pedestrian step frequency and correct speed errors during the positioning process. For heading drift, a real-time heuristic drift elimination algorithm is adopted to complete heading correction. Finally, we use landmark points with angular features observed during pedestrian walking to complete position error correction. The results of the two-dimensional indoor positioning experiment show that the endpoint positioning error is only 1.28 meters.