Abstract: High-precision real-time position provided by sensors represented by GNSS and INS is an important foundation for autonomous driving, intelligent transportation and other fields. To improve the real-time positioning performance of GNSS and MEMS INS in urban environments, a multi-system real-time PPP/MEMS INS tight integration model is constructed using real-time precise orbit and clock correction numbers broadcasted by IGS, and the positioning performance of multi-system real-time PPP and PPP/MEMS INS tight integration is analyzed and evaluated. The positioning results obtained are compared with post precise products. The experimental results show that the multi-system PPP/MEMS INS tight combination can achieve continuous and reliable positioning in the case of short GNSS interruptions, the positioning accuracy of horizontal and vertical components reaches 0.374 and 0.339 m, respectively. But there is still a gap compared with the post precise products, and the difference of 3D positioning accuracy is about 0.15 m. The velocity and attitude accuracy of the PPP/INS tight integration mainly relies on INS, and the velocity measurement accuracy is about 1 cm/s, which is not significantly different from the results obtained using the IGS post product; the observability of heading angle is weak, and the error of heading angle increases significantly during static period.

Key words: multi-system, real-time PPP, INS, tight integration, urban environment