A deterministic auto-regressive STAP approach for nonhomogenerous clutter suppression

Direct data domain (DDD) space-time adaptive processing methods avoid nonhomogenerous training samples and can effectively suppress the clutter within the test range cell. However, it suffers inevitable performance loss due to the spatial and temporal smoothing process. Furthermore, the clutter suppression ability of these methods sharply degrades when applied to non-uniform and non-linear array for airborne radar. In this paper, a novel clutter suppression approach in the direct data domain is proposed, which describes clutter characteristic of the test range cell with AR model. For convenience, the novel method is referred to as $$\hbox {D}^{3}\hbox {AR}$$D3AR. It utilizes the most system DOF. Hence, it suffers less aperture loss, compared to conventional DDD methods, e.g., the direct data domain least squares ($$\hbox {D}^{3}\hbox {LS}$$D3LS). More importantly, $$\hbox {D}^{3}\hbox {AR}$$D3AR can achieve much better clutter suupression prformance than $$\hbox {D}^{3}\hbox {LS}$$D3LS when applied to conformal array airborne radar because it does not need the spatial smoothing. The effectiveness of the $$\hbox {D}^{3}\hbox {AR}$$D3AR is verified by numerical examples for the case of a circular array.