EM-based Radar Signal Processing and Tracking

Maximizing the achievable SNR benefits the measurement’s precision that is affected by thermal noise, and track precision is enhanced given precise, unbiased, range measurements. While the radar is tracking the kinematic state (i.e., position, velocity, and acceleration) of the target, optimal signal processing requires knowledge of the target’s signature, range rate, and radial acceleration. Increasing the SNR of individual measurements by enhanced signal processing results in reduced settling time for the track filter and less valuable radar resources required to achieve a specified track quality. Moving targets introduce range-walk (RW), which degrades coherent processing leading to a reduction in SNR and a blurring of the point target response in range-Doppler (RD), resulting in degraded range estimates to the Tracker. Traditional radar architectures require excessive iterations of the track loop with the associated measurements to achieve a specified track quality threshold. In this paper, the Expectation-Maximization (EM) algorithm, an iterative and inference approach, provides improved parameter estimation and SNR by employing a Kalman filter within the Signal Processor to account for changes in the target’s kinematics. The improved state estimates are used in the Signal Processor to achieved enhanced RW correction.