Robust GNSS/INS Tightly-Coupled System with Time-Differenced Carrier Phase for Accurate MEMS-Based Navigation

Conventional global navigation satellite system (GNSS)/micro-electromechanical strapdown inertial navigation system (MEMS-SINS) tight integration encounters some major limitations in GNSS-challenged areas. It is GNSS signal impairments due to the complex environment surrounding GNSS antenna, that plagues the acquired integration performance, resulting in significant accuracy degradation of MEMS-based inertial navigation. To improve the overall performance of GNSS/MEMS-SINS tightly-coupled (TC) system, this paper proposes the following three-tier approach at different levels without either additional sensors or GNSS base stations: 1) improving the stochastic modeling for both MEMS sensors and GNSS receiver clock errors using Allan variance (AV) analysis and satellite-differencing (SD) at the dynamics level, 2) enhancing the integration by augmenting unambiguous time-differenced carrier phase (TDCP) into Kalman filter at the measurement level, and 3) eliminating the adverse effects of GNSS outliers by devising innovation-based test statistics at the information fusion level. The field test results indicate that the proposed algorithm could improve the overall navigation accuracy by about 42.14% (position), 38.96% (velocity) and 33.56% (attitude) with immunity to both GNSS outliers and cycle slips, compared with conventional TC counterpart. Moreover, it can also significantly reduce the coasting errors and improve the relative position accuracy during relatively long outages. The results demonstrate the effectiveness of the proposed algorithm in reducing the growth of MEMS-SINS error.