Comparison of filtering and smoothing algorithms for airborne radar data

The detection of ground-moving targets requires clutter cancellation, which is typically performed using space-time adaptive processing (STAP). The detections from STAP provide the measurements of range, bearing, and Doppler. These measurements can then be fed to Bayesian state estimators. In this paper, results from an airborne radar data set are processed and the performance of filtering and smoothing algorithms are compared. The standard nonlinear filtering algorithms, namely the extended Kalman filter, are used. It is found that while the smoother performance is significantly better than that of the filter, the smoothing window need not be large to obtain the superior performance.