Amplitude Information-aided Extended Target Tracking Using the PHD Filter

Radar is a commonly utilized sensor in extended target tracking (ETT). Traditional ETT methods rely primarily on spatial information (such as position and angle) from radar measurements, often leading to challenges like clustering closely spaced targets and clutter. However, modern radar systems also provide amplitude information (AI), which differs between targets and clutter. This paper proposes a novel tracking scheme that leverages AI to assist ETT. The scheme improves the measurement partitioning and the update of the extended target probability hypothesis density (ET-PHD) filter using AI. Ulti-mately, it employs Gaussian processes in conjunction with the AI-aided ET-PHD (AI-GP-PHD) filter for ETT. Simulation results demonstrate that the AI-GP-PHD filter is effective, particularly in clutter-dense environments like roadside areas in traffic scenarios.