Opportunity at Air Force Research Laboratory (AFRL)
Phenomenology-Based Adaptive Radar Signal Processing
Sensors Directorate, RY/Electromagnetics Technology Division
||Wright-Patterson AFB, OH 454337542
Research opportunities exist in physics and phenomenology-based adaptive signal processing methods for enhanced radar target detection and estimation. Classical adaptive signal processing methods for radar rely on the formation and inversion of a sample covariance matrix. However, nonstationary reflectivity properties of the scanned areas, dense target environments, and strong clutter discretes tend to introduce heterogeneities in the training data for covariance estimation. These tend to have a deleterious impact on detection and false alarm performance. Furthermore, as the dimensionality of the problem increases, the training data support for forming the covariance matrix and the computational cost of the matrix inversion are prohibitively high. To address these issues we seek to exploit a priori information from the scattering physics or phenomenology underlying a given scenario. For example, in many instances clutter can be viewed as the resultant of the scattered power from a small number of strong interference sources, thus rendering it low rank. This information can then be advantageously used to reduce the training data support and computational cost of the resulting adaptive processing algorithm. The problem of target detection is further complicated by the presence of a large number of nuisance parameters. These effects are exacerbated by systems and environmental considerations pertaining to the operational scenario. We seek novel approaches based on either a priori knowledge of the clutter scenario or on principles of invariance for this problem with a goal to maintain a constant false alarm rate and achieve robust target detection performance. Development and performance analysis of MIMO radar signal processing algorithms for the above described scenarios is of particular interest.
Adaptive radar signal processing; Physics-based methods; Constant false alarm rate; Sample support; Invariance; Robust performance; Knowledge-based methods;
Open to U.S. citizens
Open to Postdoctoral and Senior applicants