for Pre-Doppler STAP
|Edward J. Baranoski
MIT Lincoln Laboratory
244 Wood Street
Lexington, MA 02173-9108
tel: (617) 981-0480
Abstract Conventional pre-Doppler STAP algorithms perform adaptive clutter nulling for airborne radars over a portion of a coherent processing interval (CPI) and then non-adaptively combine the sub-CPI results using a Doppler constraint vector. This approach can be constructed as adaptively inflating the dimensionality of the problem followed by applying a non-adaptive constraint. Pre-Doppler algorithms have generally provided a decreased minimum detectable velocity (MDV) compared to post-Doppler algorithms since the pre-Doppler constraint is not optimal for the adapted sub-CPI outputs. Significant improvements in MDV can be obtained by allowing partial adaptivity in the constraint vector with only a fractional increase in computational complexity compared with conventional pre-Doppler techniques. This erases the performance improvement of post-Doppler over pre-Doppler algorithms. This presentation shows a subspace analysis of sub-CPI nulling which demonstrates how to minimize the computational complexity by optimally splitting the adaptivity between the clutter nulling and constraint vector optimization. This yields a multi-stage adaptive processing architecture employing low dimensionality nulling in each stage, allowing for greater parallelism and scalability.
Direct comments and questions to: email@example.com
© MIT Lincoln Laboratory. All rights reserved.