Localization error reduction for an electric aircraft tractor prototype using Kalman filter

Abstract

Operating an autonomous vehicle in an airport requires high localization precision to safely navigate the airside and avoid collisions with aircraft, buildings, and personnel. This paper presents an accuracy analysis of the localization system for a small electric aircraft tractor prototype. The system utilizes a Global Navigation Satellite System (GNSS) for positioning and enhances accuracy through data fusion with a wheel odometer using a Kalman Filter. A pilot system was installed on the autonomous small electric aircraft tractor prototype to evaluate performance. Experimental results indicate that the data fusion-based approach reduced GNSS positioning errors by 32.35% in the worst case and up to 56.47% in the best case while also increasing satellite data availability through computational estimation with an Inertial Measurement Unit (IMU). Additionally, the IMU reduces signal error data in certain areas and reduces covariance noise, resulting in more accurate and efficient movement of the electric aircraft tractor.