Robust GPS/BDS/INS tightly coupled integration with atmospheric constraints for long-range kinematic positioning

The combination of new global navigation satellite system (GNSS) has brought great benefits to reliable positioning and ambiguity resolution (AR), especially in restricted environments. However, kinematic positioning over long ranges is still a challenge due to the presence of significant atmospheric uncertainties, which contaminates the AR process. We present a tightly coupled strategy to integrate GNSS and inertial navigation system (INS) by adding ionospheric and tropospheric delay parameters and extending the stepwise AR by applying the partial ambiguity resolution (PAR) strategy. With the aid of INS predictions, the instantaneous AR can be achieved with the proposed atmospheric prediction model, along with a dual-frequency constraint ambiguity validation test. To remove the faults in both dynamic model and measurement model, a robust innovation filtering algorithm is proposed. A field vehicular test was conducted to validate the positioning performance of the proposed algorithm over long ranges. The results show that a reliable positioning solution is obtainable for the global positioning system (GPS)/BeiDou navigation satellite system (BDS)/INS integration system with baseline larger than 130 km. The average number of fixable ambiguities reaches 14.43 by applying PAR. In addition, the fixing ratio of having fixed more than three ambiguities reaches 98.57%. The results also indicate that the robust innovation filtering can efficiently detect the discrepancies in the filter.