TIGHTLY-COUPLED RTK/INS INTEGRATED NAVIGATION USING A LOW-COST GNSS RECEIVER AND A MEMS IMU

Abstract. The Global Navigation Satellite System (GNSS) real-time kinematics (RTK) is a technology to provide centimeter-level navigation services in outdoor areas. GNSS/INS integration is a typical integrated navigation system, including loosely, tightly, and ultra-tightly coupled integration, which has been studied for decades. Although GNSS aiding INS is a relatively mature system, RTK is rarely used to be integrated with INS. In this work, we present a model with a tightly-coupled (TC) scheme to integrate RTK and INS using an extended Kalman Filter (EKF). Multi-GNSS multi-frequency double-differenced pseudo-ranges and carrier phases are the updates of the measurement model. To make the ambiguity resolution, we used the well-known LAMBDA algorithm to search for the possible ambiguity series, and then use the best integer equivariant (BIE) method to decide on the most likely ambiguities with a proper weighting strategy to select from the ambiguity candidates. To test the performance of the proposed TC integration model, we implemented two field vehicular tests with a low-cost GNSS module equipped with a Micro-Electro-Mechanical System (MEMS) IMU. Using this low-cost platform, our RTK/INS integrated navigation engine can achieve centimeter-level navigation solutions under open sky conditions. In harsh environments, our TC integration system navigates on two scenarios with mean errors of 0.48 and 0.57 m, which is 51% and 28% better than a loosely-coupled system.