Research on Indoor Positioning of Multi-source Information Fusion Based on Improved Particle Filter

Abstract

In the information age, people are increasingly demanding location-based services. The traditional indoor positioning based on a single signal source has low accuracy and poor stability. On the one hand, the occlusion and echo interference in the indoor environment are serious, especially in the case of crowded environment and non-line-of-sight propagation, the UWB (Ultra Wide band) positioning accuracy will be greatly reduced. On the other hand, IMU (inertial Measurement Unit) can provide an accurate inertial navigation solution in a short time but its positioning error increases fast with time due to the cumulative error of accelerometer measurement. Therefore, in the process of pedestrian walking, we use the key information obtained by PDR (Pedestrian Dead Reckoning) to establish and update the real-time motion model of the target and calculate the prior information in the process of fusion filtering. Then, the initial position of the target is obtained through the UWB positioning solution, and the PDR positioning trajectory is corrected as the observation information in the fusion filtering process. Finally, the improved differential particle filter algorithm is used to fuse the above UWB positioning results and PDR positioning results, and make up for the advantages and disadvantages of the two positioning to improve the accuracy and stability of fusion positioning.