INTELLIGENT TIME-SPATIAL INFORMATION SERVICE TECHNOLOGY
HUANG Lu, YU Bao-guo, LI Hong-sheng, LI Jun, JIA Hao-nan, CHENG Jian-qiang, LI Ya-ning
Pseudolites have the ability to transmit the same signals as GNSS(Global Navigation Satellite System) satellites, and can provide stable and reliable positioning signals for the navigation signal obstructed environment, making it possible to achieve continuous high-precision positioning outdoors based on the existing terminal hardware conditions. Therefore, it has gradually become a research hotspot in the field of indoor positioning. In this paper, a fingerprint database matching and positioning method based on homologous multi-channel pseudolites is proposed. The variational autoencoder network that takes into account the position information is designed to learn the probability distribution characteristics of the pseudolite carrier phase information in the hidden space. Then, the mapping relationship between the hidden features of the pseudolite observation data and the indoor location is established. After this, aiming at the problem of large fluctuation of fingerprint location results, a particle filter fusion processing method is proposed to improve the stability and accuracy of the location system. In the experimental environment and airport environment, a large number of experiments verify the positioning performance of the positioning algorithm under dynamic and static conditions, and compare it with the common positioning methods based on fingerprint database matching. The results show that the dynamic average positioning accuracy is 0.39 m in the indoor test environment, and 95% of the positioning error is better than 0.85 m. In the real airport environment, the dynamic average positioning accuracy is 0.75 m, the maximum positioning error is 1.69 m, and 92% of the positioning error is better than 1m. The effectiveness of the algorithm is verified.