Influence of Pseudorange Biases on Single Epoch GNSS Integer Ambiguity Resolution

Abstract

Model strength and biases are two main factors for GNSS integer ambiguity resolution. Due to the nonideal GNSS signal, the receiver's observations will be biased. In GNSS relative positioning, the double difference pseudorange bias cannot be eliminated due to the different receiver's technical parameters between base and rover stations, which are generally ignored. This paper investigates single epoch ambiguity resolution performance in the presence of pseudorange biases. Specifically, two commonly ambiguity resolution methods are used, including the bootstrapped and the Least-Squares Ambiguity Decorrelation Adjustment (LAMBDA) methods. A transfer model of pseudorange biases affecting integer ambiguity resolution is proposed to analyze the influence of pseudorange biases theoretically. The theoretical analysis and simulation results show that the pseudorange biases can affect the float ambiguities and decrease the success rate of integer ambiguity resolution. The simulation result shows that the performance of bootstrapped method may exceed the LAMBDA method in the presence of pseudorange biases. Further, a real data experiment is conducted to confirm the influence of pseudorange biases on ambiguity resolution. Both the simulations and the experimental results show that decimeter-level pseudorange biases can also decrease the success rate of ambiguity. The real data experiment indicates that with the pseudorange biases modified, the success rate of ambiguity fixing may increase up by 5% in zero baseline. The better the geometry of the observations, the robuster the ambiguity resolution against pseudorange biases.