Objectives  Smartphones have become the mainstream terminal carrier of navigation and location services for mass users by virtue of their ubiquity, portability, and low cost. With the opening of their multi-frequency and multi-constellation global navigation satellite system (GNSS) observations, the research on high-precision positioning of smartphones has been further stimulated. However, limited by the performance of consumer-grade GNSS devices, the satellite observations of smartphones present problems such as serious signal attenuation, large pseudorange noise, and many cycle slips; and also affected by the complex urban environment, the continuity and reliability of smartphone GNSS positioning is also difficult to guarantee.

  Methods  A smartphone GNSS/MEMS (micro-electro mechanical system) integrated high-precision positioning scheme for vehicular navigation in urban conditions is proposed. Firstly, a velocity-constrained GNSS differential filtering positioning model is constructed to realize precise positioning in a general observation environment; then, through the built-in MEMS of smartphone and virtual constraints of vehicle motion, GNSS precise positioning is enhanced in challenging environments.

  Results  The experiment results show that, in open sky and tree occlusion conditions, the improved method can achieve decimeter-to-meter positioning accuracy, which is 35.2% and 78.9% higher than conventional method, respectively; in viaduct occlusion conditions, the accuracy and continuity of GNSS/MEMS fusion positioning are the best; in tunnel conditions, the cumulative position error of the MEMS mechanization is about 2.5%.

  Conclusions  The results preliminarily show that smartphone GNSS has lane-level positioning capabilities in open-sky environments, and GNSS/MEMS fusion can improve the accuracy and continuity of smartphone positioning in urban challenging environments.