Tracking Low Velocity Vehicles from Radar Measurements

Å. Göransson and B. Sohlberg (Sweden)


Source separation, nonlinear system, Digital signal processing, extended Kalman filter, frequency estimation.


In this paper an extended Kalman filter (EKF) with a smoothing window is proposed for velocity estimation of vehicles. One problem with accurate estimation is the short window length. The batch length is here limited to 2048 samples, which includes one period of a sine at the lowest frequency to bee estimated. The evaluation of the EKF is made on both computer-generated measurements with a SNR at 0 dB and real radar measurements from road vehicles. This paper considers the necessity of: the reduction of the linearisation error, the estimation of both the variance of the process noise and the measurement noise. We also take into consideration if the algorithm converges on real radar measurements.

Important Links:

Go Back