201. Fully Adaptive Radar for Variable Resolution Imaging
- Author
-
Muralidhar Rangaswamy, Graeme Smith, J. Landon Garry, Kristine L. Bell, Adam E. Mitchell, and Andrew J. Duly
- Subjects
Synthetic aperture radar ,SIMPLE (military communications protocol) ,Computer science ,Real-time computing ,ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION ,0211 other engineering and technologies ,Process (computing) ,02 engineering and technology ,law.invention ,Data set ,Autoregressive model ,law ,Radar imaging ,General Earth and Planetary Sciences ,Clutter ,Electrical and Electronic Engineering ,Radar ,021101 geological & geomatics engineering - Abstract
This paper describes the first application of the fully adaptive radar (FAR) framework for cognition to the process of radar imaging. A cognitive radar adapts to its surroundings based on its perceptions of the environment, offering improved performance for a multitude of radar applications. We implemented an autoregressive backprojection (ARBP) imaging technique for the circular synthetic aperture radar (SAR) video within the structure of the FAR framework, allowing the system to adapt its down-range and cross-range resolutions to keep the detected targets visually distinct. This simple demonstration paves the way for more advanced adaptive imaging scenarios in the future. Application of the technique to the GOTCHA volumetric SAR data set demonstrated its capability in a realistic scenario in the presence of clutter and limited target persistence. When applied to the GOTCHA data set, the adaptive imaging system’s cumulative executive optimization cost (CEOC), which is used to quantify the overall performance, was 41.3% smaller than the constant, fine resolution case. This significant improvement in CEOC comes at the expense of occasionally failing to meet imaging performance goals as the system adjusts to changes in the environment.
- Published
- 2019