Comparison and evaluation of advanced motion models for vehicle tracking 论文

摘要

Abstract—The estimation of a vehicle’s dynamic state is one of the most fundamental data fusion tasks for intelligent traffic applications. For that, motion models are applied in order to in-crease the accuracy and robustness of the estimation. This paper surveys numerous (especially curvilinear) models and compares their performance using a tracking tasks which includes the fusion of GPS and odometry data with an Unscented Kalman Filter. For evaluation purposes, a highly accurate reference trajectory has been recorded using an RTK-supported DGPS receiver. With this ground truth data, the performance of the models is evaluated in different scenarios and driving situations.