Axisymmetric extended target tracking using Gaussian process

Abstract

In extended target tracking scenarios, the size of each target is much larger than a single sensor resolution cell, which results in multiple measurements per scan. For many aerial targets, their shapes tend to be axisymmetric and in most cases the directions of motion nearly coincide with the axes of symmetry. In this paper, an extended Kalman filter with Gaussian process measurement model is used to simultaneously estimate the kinematic and extent states of such target and based on some assumptions about the target extent, a new form of the covariance function for Gaussian process is proposed. The effectiveness of the proposed method is demonstrated in several simulation experiments by comparing with the state-of-the-art algorithms, simulation results show that the proposed method has a significant improvement in estimation.