A sequential ambiguity selection strategy for partial ambiguity resolution during RTK positioning in urban areas

Usually, it is difficult to implement integer ambiguity resolution within a short amount of time for GNSS positioning in urban areas due to the contamination of non-line-of-sight signals and multipath. This study proposes a sequential ambiguity selection strategy for partial ambiguity resolution. First, the ambiguities are selected based on the filtered residuals of the phase and code measurements. In addition, the elevation angle and decorrelated variances are used as the metrics for selecting ambiguity subsets. Two kinematic experiments are carried out in urban areas to evaluate the performance of the strategy. Among the three independent strategies, the first one performs better than the others, as the dependency of observation quality on the elevation angle is low and the decorrelated variances are prone to be contaminated by biased ambiguities. When the proposed sequential ambiguity selection strategy is used, the percentage of correctly fixed epochs is increased by approximately 10–20%. The RMS of N/E/U is improved from the decimeter level (for full ambiguity resolution) to the centimeter level. The improvement is more obvious in the obstructed area and during the re-initialization phase.