Factor Graph Optimization Based Multi-epoch Ambiguity Resolution for GNSS RTK and Its Evaluation in Hong Kong Urban Canyons

With the widespread use of global navigation satellite systems (GNSS) in location-based applications, real-time kinematic (RTK) positioning has attracted much attention due to its high-precision characteristics. However, in complex environments such as urban canyons, the performance of GNSS-RTK is easily affected by insufficient and low-quality measurements, resulting in inaccurate ambiguity resolution and positioning errors of tens or even hundreds of meters. To alleviate this problem, this paper proposes a multi-epoch ambiguity resolution (AR) strategy based on factor graph optimization (FGO), and comprehensively evaluates the performance of four different AR strategies under the FGO-RTK framework. The AR strategies involved in the evaluation are: single-epoch AR without single-differenced (SD) ambiguity constraint, single-epoch AR with SD ambiguity constraint, multi-epoch AR with ambiguity merging, and our proposed method, that is, multi-epoch AR with SD ambiguity constraint. The results show that the proposed MultiSD can better utilize the temporal correlation of ambiguities across epochs, thereby improving the fixation rate and positioning accuracy in urban canyon environments. The advantages and disadvantages of the four AR strategies are validated. In short, the reliable FGO-RTK with multi-epoch AR proposed in this paper and its evaluation provide a valuable reference for the robust positioning application of GNSS in complex environments.