Long memory models for the analysis and simulation of multi-channel airborne radar measurement (MCARM) data

The performance of STAP algorithms are usually evaluated using both simulated and measured airborne radar data. Data simulation allows the effectiveness of STAP methods to be estimated on larger data sets which normally rely on an assumption of spatial homogeneity and temporal stationarity. Further, synthetic data corresponding to particular radar or terrain characteristics that contribute to degrade STAP performance (e.g. internal clutter motion (ICM), antenna array misalignment and channel mismatch) can be easily generated. Simulated clutter space-time snapshots usually rely on an assumption of identical distribution and statistical independency (11D). The aim of this paper is to analyze multi-channel airborne radar measurement (MCARM) data in order to estimate the spatial correlation of space-time snapshots in real clutter environments. Our experimental results show that clutter measured space-time steering vectors exhibit long-range dependence (LRD) characteristics. The final goal of this work is to define a reliable LRD model for strong correlated clutter space-time snapshots. An accurate characterization of clutter in multi-channel airborne/spaceborne radar data is important, because it can lead to the development of STAP algorithms with improved performance. The presented method demonstrates reliable results when applied to MCARM data files including either spatially homogeneous or nonhomogeneous clutter.