Autonomous Vehicle Pose Estimation in GNSS-Denied Areas Using Cross-Track Error Measurements

Abstract

Accurate pose estimation is critical for the safe and efficient operation of autonomous vehicles, enabling precise navigation and control in highly dynamic environments. This study presents a novel approach to vehicle pose estimation using cross-track error (CTE) measurements as supplemental inputs for GNSS denied environments. The proposed method leverages CTE, a measure of lateral deviation from a desired trajectory, to estimate the vehicle’s position and orientation in real-time using an extended Kalman filter (EKF) framework. By integrating CTE into an existing pose estimation architecture, the new system is able to provide vehicle state estimates with a greater accuracy than conventional inertial navigation systems (INS). Simulation studies were conducted on benchmark test sets, demonstrating the method’s effectiveness at reducing pose estimation errors by up to 98% in various GNSS denied situations. This research contributes to the advancement of robust and cost-efficient localization strategies, paving the way for safer autonomous vehicle navigation.