Improving B-Spline-Based 3D Extended Object Tracking Using Doppler Measurements

Object tracking is an essential task for autonomous systems to interact with the environment. Extended Object Tracking (EOT) takes advantage of the high sensor resolution to generate more than one measurement of an object and thus is able to provide the shape information along with the kinematic states. 3+1D radars (3-dimensional plus radial Doppler velocity) are suitable sensors for the tracking task, which are characterized by the radial velocity for a better kinematic estimation. In this work, we propose an improvement of an EOT algorithm to work with 3+1 D automotive radar to perform 3D EOT. In this method, the objects are modeled by extruding a side-view profile described by a B-spline curve. It further exploits the Doppler measurement to estimate the kinematic parameters by integrating the derived motion equation into the measurement model. We evaluate the methods with simulated stationary roadside radars for long-range tracking.