Dr. Hai Deng
Research Interests
Phone: (305) 348-0478
Office: EC3956
We are conducting research in a wide range of electrical engineering areas including radar sensor networks, MIMO radar, Radio Frequency Identification (RFID), adaptive signal processing, biomedical signal processing, computational electromagnetics, and VLSI design. Our recent research activities are detailed as follows:

Theory and Applications of Radar Sensor Networking Systems and MIMO Radar

Networking multiple radar systems, through information fusion at different levels, is an effective approach to significantly improving radar performances in target detection, measurement and imaging and countering the mounting radar threats such as extremely high-speed and maneuvering targets, stealth and extremely low Radar Cross Section (RCS) targets, Anti-Radiation Missile (ARM) attacks, and Electronic Counter Measures (ECM). The proposed radar sensor netowrking systems use a group of orthogonal coding radar signals, allowing the multiple radar stations in the system to operate in both monostatic and bistatic modes. The specific areas that are investigated include: (a) the efficient optimization algorithms for the design of orthogonal coding radar signals, which are the key to the feasibility of the radar sensor networking systems; (b) the optimum and suboptimum signal detection schemes; (c) the effective algorithms for highly accurate target information measurements and target recognition; (d) the innovative schemes using the systems for effective target recognition of high-speed maneuvering targets such as Intercontinental Ballistic Missiles (ICBM); and (e) the optimum radar network architectures and configurations for various functions and geometric scenarios.

Multiple-Input Multiple-Output (MIMO) radar is very similar to the radar networks in terms of strucutre and operations. The multiple transmitters and receivers of MIMO radar might be geograhically close and coherent siganl processing is more achievable; while those of radar networks may be spatially more separated and the overall performance is improved through diversity.

Design and Applications of Radio Frequency Identification (RFID) Systems

Radio Frequency Identification (RFID) has been widely used in a variety of areas such as logistics and supply chain, document tracking, access control, and wireless commerce. Our work in this area is devoted to improving RFID performance and extending its applications. Specifically, we are trying to design and develop RFID tags with strong sensing capabilities and with ultra-long operational ranges. We are also developing innovative techniques to improve RFID detection capability in noise/clutter environment and to reduce interference from other readers when multiple RFID systems operating at the same location. In addition, we are interested in developing and applying RFID systems for anti-counterfeit and theft-prevention

Space-Time Adatpive Processing

Airborne or space-based sensor systems (such as microwave radars) need to operate in difficult environments where the detection of small, highly-maneuvering targets against strong clutter background in the presence of jamming. Space-Time Adaptive Processing (STAP) is an effective way to suppress both clutters and jamming signals simultaneously. However, there are some serious challenges to implementing STAP algorithms in airborne sensor systems. The major challenges are how to accurately estimate the statistical characteristics of inhomogeneous ground clutters and how to effectively reduce STAP computational complexity without degrading system performances. The research in this area is to develop innovative techniques to overcome the difficulties in applying STAP.

High-Speed VLSI Circuit Design, Modeling and Simulation

With increasingly high integration and high-speed operation for VLSI circuits, the obstacles to the next generation VLSI design will boil down to power consumption and on-chip interconnect resistance and inductance. Our research work in this area concentrates on the design of novel high-speed circuits with low-power consumption feature, such as Domino circuits, self-timed circuits and nanowire circuits, and on-chip interconnect modeling and simulation.

Fast Computational Electromagnetic Techniques and Applications

Computational Electromagnetics (CEM) is crucial for antenna design, performance prediction for radar and wireless scattering, and microwave and high-speed VLSI circuit modeling and simulation. But the prohibitive computational complexity of the CEM techniques such as Method of Moments prevents widespread applications of CEM techniques to large-scale electromagnetic problems. The research in this area aims to seek innovative approaches to fundamentally reducing the complexity of computational electromagnetic techniques. The specific research areas include: (a) developing fast wavelet-based method of moments. The new method is developed to overcome the N square complexity and memory bottlenecks existing in the current fast transform methods; and (b) developing fast preconditioning methods for iterative computational electromagnetic solvers such as Fast Multiple Method (FMM).