Staring RF Signal Processing Challenges Linda J. Moore * Jason T. Parker * LeRoy A. Gorham * Uttam K. Majumder * Michael J. Minardi * Steven M. Scarborough * Abstract — Traditionally, distinct radar modes have been employed to accomplish specific tasks such as imaging an area of interest, or detecting and track- ing moving targets. Staring circular synthetic aper- ture radar (S-CSAR) provides unique opportunities for exploitation of radio frequency (RF) data col- lected over a large ground spot. The same phase history may be processed in different manners to generate simultaneous S-CSAR products such as 2- D Video SAR, coherent and non-coherent change detection (CCD and NCD), and ground moving tar- get indication (GMTI). Advanced signal processing techniques can take advantage of the S-CSAR ge- ometry to produce 3-D scene reconstructions. The ability to transmit, record and process large volumes of S-CSAR data, to create high fidelity exploitation products, in real-time, poses significant challenges. This paper addresses several open problems in this research area. 1 INTRODUCTION Air-to-ground radars have traditionally been devel- oped as narrow beam, scanning systems. These systems have significant constraints such as low re- visit rates, low power efficiency, low resolution, and difficult sensor management problems. Also, since scanning systems are pre-tasked to perform specific functions, forensic analysis of old data is often im- possible. Staring circular synthetic aperture radar (S- CSAR) offers significant improvements in exploita- tion functions by continually staring at an area of interest (AOI) with a very wide beam. The sensor platform flies a circular flight path at a constant elevation around a fixed scene center, and the AOI is constantly radiated with SAR-like pulses. This type of system largely overcomes the limitations of the traditional air-to-ground approach. The primary benefit of S-CSAR systems is that several different functions can be performed si- multaneously with the same radio frequency (RF) data, such as imaging, stationary target detec- tion, ground moving target indication (GMTI), and change detection. Notably, the choice of data prod- uct can be made after the collection, thus simpli- fying sensor management and allowing for forensic analysis. Several challenges exist in terms of generating * Sensors Directorate, US Air Force Research Laboratory, 2241 Avionics Circle, Building 620, Wright-Patterson AFB, OH 45433, e-mail: linda.moore2@wpafb.af.mil high fidelity exploitation products from S-CSAR data collected over large ground spots. Challenge S-CSAR data sets and several open problems in the research areas of Video SAR, change detection, GMTI and 3-D scene reconstruction are offered in this paper. 2 CHALLENGE DATA SETS AFRL has produced three carefully documented challenge problems for use by the larger research community. The challenge problems focus on 2- D/3-D image reconstruction [1]; SAR GMTI [2]; and change detection and lossy phase history com- pression [3]. 3 VIDEO SAR Video SAR is a sequence of SAR images of the en- tire AOI produced at a given resolution and frame rate. Each SAR image is formed to the same pixel locations on the ground, and the height of each pixel could be derived from a Digital Eleva- tion Model (DEM). Through this process, the im- ages must be orthorectified so that the scene does not appear to rotate as the platform flies its circu- lar path. Since it takes several seconds to form a synthetic aperture to achieve equivalent resolution in cross-range as that provided via bandwidth in range, successive frames of Video SAR have over- lapped apertures. Backprojection (BP), see [4] for a tutorial and sample code, was selected over the more compu- tationally efficient Polar Format Algorithm (PFA) to leverage its inherent flexibility and pulse-wise processing, which allows intermediate results to be reused when processing overlapping apertures. To overcome the order N 3 computational requirements of BP and achieve real-time video SAR, fast BP al- gorithms have been developed [5, 6]. Improved computational efficiency for Video SAR remains an open research problem. S-CSAR introduces several processing challenges for an im- age formation processor (IFP). The relatively large scene invalidates a widely used plane wave approx- imation, which results in defocused and distorted images if the range curvature is not addressed. The IFP must also account for varying terrain and the non-linear circular flight path. Also, the overlapped U.S. Government work not protected by U.S. copyright 279