Small cycle slip detection using singular spectrum analysis

In Global navigation Satellite System (GNSS) based positioning, the use of carrier phase measurements is widening day by day due to their preciseness as compared to code delay measurements. Although, carrier phase measurements are precise but they suffer from anomalies such as cycle slips and receiver clock jumps in addition to other error sources such as satellite-user dynamics and atmospheric delays. On one end, the detection and exclusion of these anomalies is critical for accurate and reliable positioning. On the other end, it is very difficult to detect these anomalies especially in dynamic environment due to irregular user dynamics. We propose a novel algorithm for separating and localizing these anomalies from the satellite-user dynamics. The proposed approach is based on extracting the singular spectrum of windowed carrier phase measurements. An optimal choice of different parameters ensures that the extracted singular spectrum is affected only by anomalies such as cycle slips and is independent of satellite-user dynamics. Simulation results, supported by real GNSS data analysis, indicate improved accuracy and enhanced robustness against such anomalies with respect to traditional approach using optimal time differencing.