Ozlem Kilic

Parallelizing Fast Multipole Method for Large-Scale Electromagnetic Problems Using GPU Clusters

This letter investigates the solution of large-scale electromagnetic problems by using the single-level Fast Multipole Method (FMM). Problems of large scale require high computational capability that cannot be accommodated using conventional computing systems. We investigate a parallel implementation of FMM on a 13-node graphics processing unit (GPU) cluster that populates Nvidia Tesla M2090 GPUs. The implementation details and the performance achievements in terms of accuracy, speedup, and scalability are discussed. The experimental results demonstrate that our FMM implementation on GPUs is much faster than (up to 700 ×) that of the CPU implementation. Moreover, the scalability of the GPU implementation is very close to the theoretical linear expectations.

Analysis of Moving Human Micro-Doppler Signature in Forest Environments

Automatic detection of human motion is important for security and surveillance applications. Compared to other sensors, radar sensors present advantages for human motion detection and identiflcation because of their all-weather and day-and-night capabilities, as well as the fact that they detect targets at a long range. This is particularly advantageous in the case of remote and highly cluttered radar scenes. The objective of this paper is to investigate human motion in highly cluttered forest medium to observe the characteristics of the received Doppler signature from the scene. For this purpose we attempt to develop an accurate model accounting for the key contributions to the Doppler signature for the human motion in a forest environment. Analytical techniques are combined with full wave numerical methods such as Method of Moments (MoM) enhanced with Fast Multipole Method (FMM) to achieve a realistic representation of the signature from the scene. Mutual interactions between the forest and the human as well as the attenuation due to the vegetation are accounted for. Due to the large problem size, parallel programming techniques that utilize a Graphics Processing Unit (GPU) based cluster are used.

Joint DoA-range-Doppler tracking of moving targets based on compressive sensing

In this paper, the tracking of moving targets using a stepped frequency linear antenna array is addressed through the compressive sensing (CS) framework. Multiple targets are resolved in the three-dimensional directions of arrival (DoA)-range-Doppler space for each time instant. The Orthogonal Matching Pursuit (OMP) algorithm is utilized to reconstruct the target space using measurements at a small number of randomly selected frequencies. The performance of the proposed approach is evaluated through simulation result in terms of accuracy and computational complexity.

APPLICATION OF ARTIFICIAL IMMUNE SYSTEM ALGORITHM TO ELECTROMAGNETICS PROBLEMS

This paper investigates the use of clonal selection principles based on our immune system for optimization applications in electromagnetics. This concept is based on our immune systems ability to respond to an antigen and produce a pool of anti-body secreting cells. In addition to the common implementations of this algorithm where the a-nity maturation and cloning principles of clonal selection principles are used, we utilize memory and the cross-over concepts that are common to other bio-inspired methods. The performance of the algorithm is investigated for well known mathematical test functions and its potential is demonstrated in the context of the design of a radar absorbing material and a planar phased array antenna with speciflc radiation and null characteristics. Classical optimization techniques typically require an initial estimate reasonably close to the flnal result in order to avoid stagnation at a local optimum point. They also tend to be computationally intensive as they require analytical calculations such as derivatives. Nature provides heuristic optimization methods that rely on the techniques devised by difierent species over thousands of years for survival and can be utilized in engineering applications. Some of these optimizations are utilized for survival by difierent species. These approaches tend to be agent based as they can simultaneously sample the optimization space for a certain number of randomly inspired possibilities, with each iteration adding more intelligence to the heuristic search steps involved. Some examples of nature inspired optimization techniques are the genetic algorithm (GA) (1), particle swarm optimization (PSO) (2), ant colony optimization (ACO) (3), and the artiflcial immune system (AIS) (4), which is the main focus of this paper.

Interference analysis for spot beam partitioning in cellular satellite communication systems

It is observed that the sub-beam approach will reduce the gain requirements by 3 dB and 3.6 dB or partition factors of four and seven, respectively. This implies more than 50% reduction in the aperture size. The interference study shows that the partition factor of seven results in high levels of interference as the sub-beams are defined at only 0.4 dB down relative to the peak, and the half power beamwidth is quite broad so that spatial isolation can not work favorably to reduce interference levels. However, the partition factor of four produces similar interference levels as the spot beam configuration (26.92 versus 26.87) using only half of the aperture size. The investigation shows that the spot beam partitioning is a good technique to reduce the antenna size while maintaining the system requirements, such as edge gain and center to center interference.

System aspects and transmission impairments of active phased arrays for satellite communications

The communications link and system aspects of active phased arrays that are used in multiple-beam satellite systems are assessed through measurements and analysis. Three link parameters are investigated and their effects on the overall carrier-to-interference ratio (CIR) are quantified. The first parameter is the intermodulation components that are generated at the nonlinear amplifier outputs and contribute to well-formed interference in the far-field radiation of the array. The second is the bit-error ratio (BER) degradation due to the multi-carrier operation of the active array. Measurement results are shown to demonstrate this effect. The third link parameter is the cochannel interference caused by frequency reuse in multiple-beam systems. The paper starts by reviewing early developments of phased arrays for multiple-beam satellite communications applications. A key component in these developments is the modular monolithic microwave integrated circuit (MMIC) beam-forming matrices that generate a number of simultaneous and independently digitally controlled beams

Phase-Based Methods for Heart Rate Detection Using UWB Impulse Doppler Radar

Ultra-wideband (UWB) pulse Doppler radars can be used for noncontact vital signs monitoring of more than one subject. However, their detected signals typically have low signal-to-noise ratio (SNR) causing significant heart rate (HR) detection errors, as the spurious harmonics of respiration signals and mixed products of respiration and heartbeat signals (that can be relatively higher than heartbeat signals) corrupt conventional fast Fourier transform spectrograms. In this paper, we extend the complex signal demodulation (CSD) and arctangent demodulation (AD) techniques previously used for accurately detecting the phase variations of reflected signals of continuous wave radars to UWB pulse radars as well. These detection techniques reduce the impact of the interfering harmonic signals, thus improving the SNR of the detected vital sign signals. To further enhance the accuracy of the HR estimation, a recently developed state-space method has been successfully combined with CSD and AD techniques and over 10 dB improvements in SNR is demonstrated. The implementation of these various detection techniques has been experimentally investigated and full error and SNR analysis of the HR detection are presented.

GPU Cluster Implementation of FMM-FFT for Large-Scale Electromagnetic Problems

The fast multipole method (FMM) combined with fast Fourier transform (FFT) is investigated for the solution of large-scale electromagnetic problems, which require high computational capability that cannot be accommodated using conventional computing systems. The implementation is parallelized on a 13-node graphics processing unit (GPU) cluster that populates Nvidia Tesla M2090 GPUs. The experimental results based on our FMM-FFT implementation on GPUs demonstrate up to 957 times speedup compared to that of the single-core, single-node CPU implementation. The implementation details and the performance achievements in terms of accuracy, speedup, and scalability are discussed.

Reaping the Processing Potential of FPGA on Double-Precision Floating-Point Operations: An Eigenvalue Solver Case Study

Many scientific applications such as electromagnatics require their operations carried out in double-precision floating-point format. The efficiency of these applications is mainly subject to the floating-point processing performance on the target processors. In this work, we use an eigenvalue solver application as a case study to demonstrate the processing potential of an FPGA device when dealing with floating-point operations. Relying on deep pipelines and large local memory directly accessible within the FPGA device, more than 20

Compact Rotman Lens Structure Configurations to Support Millimeter Wave Devices

\times