Enhancing GNSS/INS Integration Performance with B-Spline Trajectory Optimization for Vehicle Navigation in Urban Environments

Precise positioning is essential for autonomous navigation in complex urban scenarios. However, Global Navigation Satellite System (GNSS) signal interruptions caused by urban buildings and infrastructure result in significant positioning errors for GNSS/Inertial Navigation System (INS) integrated systems. This paper proposes a B-spline trajectory optimization (BSO) method to enhance positioning accuracy during GNSS outages. The method employs weighted B-spline representation where control points are uniformly extracted from Extended Kalman Filter position estimates. The B-spline representation inherently constrains the trajectory shape during outages, and an uncertainty-based weighting strategy amplifies this constraint effect by emphasizing high-certainty control points. Experimental validation using real-world urban datasets demonstrates the effectiveness of the proposed method. For five representative outage segments, BSO achieves RMSE reductions in horizontal positioning, with improvements ranging from 5.9% to 17.8%. Error evolution analysis reveals that BSO effectively constrains high-uncertainty trajectory segments through the weighted Bspline formulation. The proposed method provides an effective approach for enhancing GNSS/INS positioning accuracy and reliability in urban environments.