Undifferenced and Uncombined ionospheric-weighted PPP-RTK/INS tightly integration model

PPP-RICK, otherwise known as integer ambiguity resolution enabled precise point positioning, its implementation requires a network consisting of multiple receivers to generate corrections, and a user equipped with a standalone receiver to be positioned. In case of a medium-scale network with inter-station distances of a few hundreds of kilometers, it is sensible to turn to the ionosphereweighted (IW) PPP-RICK, in which, GNSS observation equations formulated at undifferenced and uncombined level have been further strengthened by imposing constraints on the stochastic characteristics of the ionosphere, besides making the model compatible for multi-frequency and single-frequency users. Nevertheless, the vulnerability of satellite signals severely limits the application in challenging environments. In this contribution, we integrate the IW-PPP-RTK with the inertial navigation system (INS) through setting up a tightly-coupled (TC) model with dual- and single-frequency observations. We evaluate the performance of this TC model with two sets of vehicle-based data collected in GNSS-deprived environments that are approximately 70 km from the nearest reference station. The results indicate that, the positioning accuracy of dual frequency PPP-RTK can be improved from decimeter level to centimeter level in a semi-urban environment with INS augmentation, and the refixed time is reduced from 11 s to 3. 6 s in a typical signal interruption scenario. Moreover, in semi-urban and urban environments, the availability of dual-frequency PPP-RTK/INS TC is 100% and 96. 9%, respectively, when the horizontal error is less than 0. 2 m. Besides, taking the lane-level navigation with single-frequency PPP-RTK/INS TC into consideration, evaluating the availability that the horizontal error less than 0. 5 m, the single-frequency PPP-RTK/INS TC are 98. 97% and 76. 1% in semi-urban and urban environments, respectively.