Abstract :
Time-dependent Fourier transform plays an important role in radar signal analysis. The Doppler frequency carries information about the relative velocity of a moving target regarding the radar antenna. However, because the target movement may be very complex, with temporary high-intensity manoeuvres, the Doppler frequency must be estimated by using a window function. In order to minimise the estimation error, the window function width must be dependent on the target manoeuvre as well as on the signal-to-noise ratio. Here, a system consisting of a Doppler filter and a second-order Kalman filter, connected in an adaptive structure, has been proposed as an efficient solution to this problem. The proposed approach enables the window width adaptation, based on the estimates of target acceleration and signal-to-noise ratio, generated by an adaptive Kalman filter. Simulation results demonstrate the advantage of this structure compared to the fixed-width, window-based algorithms, giving acceptable results even for very bad signal-to-noise conditions.
