A Comparison of Kalman Filter-based Approaches for Elliptic Extended Object Tracking

In this work, we discuss and compare Kalman filter-based approaches for tracking an elliptic extended object parameterized with orientation and semi-axes lengths. The methods include an Extended Kalman filter (EKF) implementation of the Random Hypersurface Model (RHM) approach using a radial function, an EKF-based approach for the Multiplicative Error Model (MEM), called MEM-EKF*, and a method for tracking the semi-axes independently, the Independent Axes Estimation (IAE) approach. We discuss pros and cons of the methods and compare them in various scenarios with a maneuvering object.