Improving Low-Cost GNSS Positioning Through Hybrid-Frequency PPP Ambiguity Resolution With Classified Augmented Regional Corrections

High-precision positioning and navigation using low-cost global navigation satellite system (GNSS) chipsets have profound potential for mass-market applications such as personal mapping, and smartphone-based or vehicle-based location service. Particularly, the emergence of consumer-grade GNSS chipsets with dual-frequency (DF) observations available, as well as the development of multiconstellation GNSS, further promotes the popularization and diversification of high-precision GNSS applications. In this study, a regional network augmented precise point positioning (PPP) ambiguity resolution (AR) model was developed and extended to include several hybrid-frequency cases such as frequency bands containing, intersecting, or being disjoint. Besides, an observation classification approach was proposed to further enhance the performance of PPP AR by categorizing observations based on the fixing status of regional corrections. Real-world vehicle kinematic experiments were conducted using raw observations from an ublox-F9p receiver and an android smartphone (Xiaomi-Mi8) connected to an external antenna to verify the proposed model. Results indicate that for the Ublox F9p receiver, the hybrid-frequency PPP AR can achieve 1–4 cm in the horizontal components with a fixing rate over 96% and the time to first fix (TTFF) of 29 s in kinematic scenarios, exhibiting significant improvement compared to the traditional PPP solutions with the accuracy of decimeter to meter level. In addition, the observation classification approach shows significant advantages in the Xiaomi-Mi8 test where the availability of positioning accuracy better than 0.1 m is improved from 68.3% to 88.1%.