Key points are not available for this paper at this time.
As Global Navigation Satellite Systems (GNSS) and inertial measurement units (IMUs) are gradually emerging as ubiquitous, utilizing redundant sensors to achieve accurate and robust large-scale navigation holds great promise. In this article, we propose a tightly coupled precise point positioning-real-time kinematic (PPP-RTK) /IMU system with multiple homogeneous sensors to enable high-precision and high-availability vehicle navigation. In it, a stacked Kalman filter is employed to fuse raw pseudorange and carrier phase measurements from all GNSS terminals, driven by a core IMU designated randomly. In PPP-RTK processing, precise atmospheric and bias corrections from a GNSS server are leveraged to quickly resolve carrier phase ambiguity, thus ensuring centimeter-level positioning. Besides, rotation and translation constraints from redundant IMUs are imposed on the estimation pipeline to further improve the state estimation. Real-world experiments show that the proposed method can achieve an availability of 98. 7% (horizontal position error 0. 1 m) in a half-open-sky case, which is far better than the state-of-the-art methods. Also, redundant sensors not only contribute significantly to the integer ambiguity resolution, but are also able to promise continuous state estimation in the case of sensor failure.
Shen et al. (Fri,) studied this question.