Adaptive Calibration of Parameters for Accurate Vehicle Localization

초록

We introduce an innovative method for enhancing localization accuracy using a new, simplified 2-DOF bicycle model designed to better reflect real-world dynamics. The model incorporates two key parameters that are adaptively calibrated to align with real conditions through sensor data. Our approach dynamically adjusts vehicle parameters to more accurately mirror real-world behavior, leading to improve localization accuracy. Extensive testing across varied racing scenarios demonstrates that this method provides notable improvements in localization precision and robustness. These findings underscore the potential of our adaptive calibration framework to enhance localization in autonomous vehicles, particularly in challenging and dynamic environments.

키워드