Physical-Knowledge Based Radar Signal Processing
Applied signal processing for radar and radios are critical for safety, hazard detection and clutter mitigations. There are two approaches have been used: (1) Fully adaptive processing based on a large amount of training data, which is fundamentally a phenomenological or empirical approach, (2) Detailed physical modeling, which tends to be very complicated. The physical-knowledge based processing, on the other hand, combines these two approaches and achieves an optimal balance among complexity and efficiency. Physical knowledge derived from Monte-Carlo simulation, modeling, lab model measurements, and actual data are now combined in a unified framework as knowledge model, and I have developed a suite of methods to utilize this knowledge in adaptive filtering, classification and integrated decision making. This innovative concept has been successfully used in the projects such as:
- Polarimetric diversified ground and airborne radar sensing of hydrometeor hazards (heavy rain, hail, snow and mixtures).
- Airborne wind (and turbulence) hazard detection and estimation.
- Wind-Turbine Radar Clutter (WTRC) characterization and mitigation.
- Anti-interference GPS receiver in complex EM environemnt.
- Hard target (such as air-traffic) detection within weather clutter.
The currentfocus of this program is to improve the existing Gaussian Mixture Models (GMM) and lab measurement sensitivities as well as accuracies, and innovative filtering and classification algorithm development on real-time platforms.
Space, Weight and Power (SWaP) Constrained Radar Sensors
Extreme requirements from Unmanned Systems such as UAV and satellites demand small, integrated and multi-functional radar sensor systems. In this area, this team emphasizes on a system solution combining both software and hardware. The ultra-fast system chain, surface-mount integrated microwave module (IMM) and assemblies, real-time RapidIO network-based embedded DSP, and ultra-resolution target imaging are the key technologies being developed in this area. Example projects ongoing in this area include:
- TR module design and prototyping for Multi-functional Phased Array (MPAR).
- IMM design and assembly for commerical weather radar development.
- Passive radar with RF energy harvesting capability.
- Ultra-fast retrodirective noise radar transceiver.
- Ultra-low power radar sensor network for traffic sensing and control at high-way workzones.
- Sparse reflectarray antenna and phase-only pattern optimization.
- RapidIO-based DSP-FPGA farm system and real-time radar algorithm-testbed.
- Decorrelated-Scan MUSIC algorithm for super-resolution AOA estimations.
The focus of this area for the near term is supporting various high-end testbed system for mobile and airborne radar research and data processing.
RF Transceiver Optimization for Airborne and Space-Borne Radar
The incorporation of solid-state amplifiers in modern radar systems brings about new challenges in nonlinearity effects and pattern artifacts. The integrated systems can achieve the claimed performances only after the effects of such distortions are removed. Innovations in both software and hardware are being made from this team, as examples:
- Adaptive 2D MMSE radar receiver to remove both range sidelobe and antenna sidelobe for observing both hard and distributed targets.
- Adaptive pre-distortion based on power amplifier (PA) characterization and behavior modeling.
- Implementation of adaptive and high-efficiency PA power control.
- Waveform optimization for combined radar and communication (e.g., PM-PPM waveform optimization for ADS-B collision avoidance radar).
The focus of this area currently is real-time adaptive transceiver implementations.