Parallel matrix inversion techniques

Lau, Kong Cheen; Kumar, Mohan; Venkatesh, R.

doi:10.1109/icapp.1996.562917

Cited by 7 publications

(2 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One of the major contributors to execution time in MVDR is a matrix inversion that must be performed upon the CSDM with each iteration of beamforming. Many fine-grained algorithms exist for parallel matrix inversion [14][15][16][17][18][19] . However, like systolic array algorithms, these algorithms generally require too much communication to be feasible on a distributed system.…”

Section: Introductionmentioning

confidence: 99%

Distributed Parallel Processing Techniques for Adaptive Sonar Beamforming

et al. 2002

View full text Add to dashboard Cite

Quiet submarine threats and high clutter in the littoral environment increase computation and communication demands on beamforming arrays, particularly for applications that require in-array autonomous operation. By coupling each transducer node in a distributed array with a microprocessor, and networking them together, embedded parallel processing for adaptive beamformers can glean advantages in execution speed, fault tolerance, scalability, power, and cost. In this paper, a novel set of techniques for the parallelization of adaptive beamforming algorithms is introduced for in-array sonar signal processing. A narrowband, unconstrained, Minimum Variance Distortionless Response (MVDR) beamformer is used as a baseline to investigate the efficiency and effectiveness of this method in an experimental fashion. Performance results are also included, among them execution times, parallel efficiencies, and memory requirements, using a distributed system testbed comprised of a cluster of workstations connected by a conventional network.

show abstract

Section: Introductionmentioning

confidence: 99%

Distributed Parallel Processing Techniques for Adaptive Sonar Beamforming

et al. 2002

View full text Add to dashboard Cite

show abstract

“…Various methods of matrix inversion are described in the literature, [27][28][29] and much work exists in the area of parallel matrix inversion. [30][31][32][33][34][35] However, these are fine-grained methods involving frequent inter-processor communication, and are thus not suitable for distributed processing systems which typically have a large communication overhead. George et al 26 used a matrix inversion technique based on Gauss-Jordan elimination with full pivoting.…”

Section: Introductionmentioning

confidence: 99%

Parallel Algorithms for Robust Broadband MVDR Beamforming

2002

View full text Add to dashboard Cite

Rapid advancements in adaptive sonar beamforming algorithms have greatly increased the computation and communication demands on beamforming arrays, particularly for applications that require in-array autonomous operation. By coupling each transducer node in a distributed array with a microprocessor, and networking them together, embedded parallel processing for adaptive beamformers can significantly reduce execution time, power consumption and cost, and increase scalability and dependability. In this paper, the basic narrowband Minimum Variance Distortionless Response (MVDR) beamformer is enhanced by incorporating broadband processing, a technique to enhance the robustness of the algorithm, and speedup of the matrix inversion task using sequential regression. Using this Robust Broadband MVDR (RB-MVDR) algorithm as a sequential baseline, two novel parallel algorithms are developed and analyzed. Performance results are included, among them execution time, scaled speedup, parallel efficiency, result latency and memory utilization. The testbed used is a distributed system comprised of a cluster of personal computers connected by a conventional network. 69 J. Comp. Acous. 2002.10:69-96. Downloaded from www.worldscientific.com by UNIVERSITY OF CALIFORNIA @ SAN DIEGO on 04/12/15. For personal use only. 70 P. Sinha, A. D. George & K. Kim other sources and sensor noise and isolate the true source directions accurately. There are various ABF algorithms proposed in the literature. 3-7 Castedo et al. 4proposed a method based on gradient-based iterative search that makes use of properties of cyclostationary signals. Krolik et al. 5 defined a steered covariance matrix as an efficient space-time statistic for high-resolution bearing estimation in broadband settings. The algorithm considered here is the MVDR algorithm described by Cox et al. 6 MVDR is an optimal technique that selects the weights in such a way that the output power is minimized, subject to the constraint that the gain in the steering direction being considered is unity. This optimization of the weights leads to effective interference-rejection capabilities, as nulls are steered in directions of strong interference. Wax and Anu 8 presented an analysis of the signal-to-interferenceplus-noise ratio (SINR) at the output of the minimum variance beamformer. Raghunath and Reddy 9 analyzed the finite-data performance of the MVDR beamformer. Zoltowski 10 developed expressions that describe the output of the MVDR beamformer in the presence of multiple correlated interfering signals. Harmanci et al. 11 derived a maximum likelihood spatial estimation method that is closely related to minimum variance beamforming.However, classical MVDR beamforming techniques suffer from problems of signal suppression in the presence of errors such as uncertainty in the look direction and array perturbations. To mitigate such problems, various constrained MVDR beamforming schemes have been proposed in the literature. These schemes enhance the robustness of the MVDR algorithm. White noise inj...

show abstract

Merging, sorting and matrix operations on the SOME-Bus multiprocessor architecture

Katsinis

2004

Future Generation Computer Systems

View full text Add to dashboard Cite

Parallel matrix inversion techniques

Cited by 7 publications

References 7 publications

Distributed Parallel Processing Techniques for Adaptive Sonar Beamforming

Distributed Parallel Processing Techniques for Adaptive Sonar Beamforming

Parallel Algorithms for Robust Broadband MVDR Beamforming

Merging, sorting and matrix operations on the SOME-Bus multiprocessor architecture

Contact Info

Product

Resources

About