A cross-epoch ambiguity tracking method for GNSS/INS/Vision-based urban navigation

Abstract High-precision Global Navigation Satellite System (GNSS) positioning in urban environments is frequently degraded by carrier-phase interruptions and cycle slips, leading to repeated ambiguity reinitialization and prolonged reconvergence after signal reacquisition. To address this challenge, a Cross-Epoch Ambiguity Tracking (CEAT) method is proposed for tightly coupled GNSS/Inertial Navigation System (INS)/Vision navigation. The converged ambiguities are preserved across discontinuous carrier-phase arcs and reused after signal reacquisition for cycle-slip detection and ambiguity rectification. Carrier-phase residuals are constructed using the stored ambiguities and propagated pose, enabling residual-based cycle-slip detection and Least-squares AMBiguity Decorrelation Adjustment (LAMBDA)-based integer estimation. The affected ambiguities and their corresponding uncertainties are then rectified according to the resolved cycle slips, thereby accelerating ambiguity convergence and improving positioning performance. The proposed method is evaluated using simulated cycle-slip tests and real-world urban and campus datasets. Comparisons with representative cycle-slip detection methods and sliding ambiguity window frameworks demonstrate the effectiveness of CEAT in cycle-slip detection, ambiguity rectification, and positioning performance. Ablation studies and prior-position sensitivity experiments are further conducted to comprehensively assess the reliability and practical applicability of the method.