Performance Enhancement of Tightly Coupled GNSS/IMU Integration Based on Factor Graph With Robust TDCP Loop Closure
Shiyu Bai, Jizhou Lai, Pin Lyu, Yiting Cen, Xin Sun, Bingqing Wang
Abstract
To autonomous systems (ASs) equipped with single-frequency GNSS chipsets, it is hard to fully use the high-precision carrier phase to assist positioning as the inherent integer ambiguity cannot be reliably obtained using only single-frequency observations, especially in obstructed environments. Although the time difference can be done to eliminate the integer ambiguity to form accurate constraint in graph based fusion, cycle slip is not properly handled in current methods, leading to degraded performance. In this paper, a performance enhancement of tightly coupled GNSS/IMU integration based on factor graph with robust time-differenced carrier phase (TDCP) loop closure is proposed. Different from traditional methods, cycle slip detection is firstly performed, and the effect of cycle slip can be eliminated in the formation of loop closure, which can bring about better robustness. Doppler frequency shift is used to detect cycle slip and carrier phase measurements at corresponding epochs are then selected to build TDCP loop closure. The proposed robust TDCP loop closure is fused with pseudorange, Doppler frequency shift and IMU via factor graph, which can make the full use of the absolute and relative measurements to enhance the performance of tightly coupled GNSS/IMU integration. Simulation and field tests are conducted to validate the proposed method. The results show that the proposed method can improve the positioning performance with robust TDCP loop closure. Meanwhile, the computational load of proposed method is discussed.