Sub-Modularity and Antenna Selection in MIMO Systems

Vaze, Rahul; Ganapathy, Harish

doi:10.1109/lcomm.2012.070512.120912

Cited by 34 publications

(25 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Submodularity enables the use of a "greedy selection" algorithm that performs this optimization to within a constant factor of the global maximum, while requiring a significantly lower number of computations. In wireless MIMO systems, it is known that the problem of selecting a subset of antennas at the receiver for signal detection is submodular [28]. In the MMF case, the link rate is observed to be submodular with respect to the grid of detectors chosen, while keeping the laser configurations fixed.…”

Section: B Submodular Searchmentioning

confidence: 99%

“…The greedy search approach aggressively seeks to place detectors where the dominant modes' power is received. While the efficient greedy search technique is useful for developing static detector arrays, a small number of computations can also be used to enable dynamically reconfigurable detector arrays for reduced complexity signal processing [28].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Analysis of Laser and Detector Placement in Incoherent MIMO Multimode Fiber Systems

Appaiah¹,

Zisman²,

Das³

et al. 2014

J. Opt. Commun. Netw.

View full text Add to dashboard Cite

Abstract-Conventional large-core multimode fibers (MMFs) are preferred for use in short to medium haul optical fiber links, owing to their tolerance to misalignment and low deployment costs; however, data rates through MMFs are limited by modal dispersion. Digital signal processing with multiple-input multiple-output (MIMO) techniques has offered promising solutions to overcome the dispersion limitations of MMFs, but the impact of the geometry of laser and detector arrays on the achievable data rate is not established. To this end, we use a field-propagation-based model to gauge the impact the geometry of lasers and detectors can have on the achievable ergodic and outage rates of incoherent MIMO-MMF links. Laser and detector array geometries were investigated using a grid-based method to optimize the positions of lasers and detectors for a 1 km MIMO-MMF link. Simulations reveal that systems with appropriately designed laser/detector geometries could improve the achievable rate over the fiber by more than 200% over random laser/detector arrays. The grid-based search technique, however, is limited due to high computational requirements for fine grids. As an alternative, we developed a suboptimal "greedy" selection approach to design detector geometries, which produces detector geometries that attain more than 90% of the rate obtained with an exhaustive search, while requiring less than 0.2% of the computation. The low computation requirements and high performance of the greedy selection approach also motivate the use of dynamically reconfigurable detector arrays to achieve high data rates with reduced signal processing complexity. Methods are also presented for clustering detector elements to obtain more consolidated segmented detectors with better fill factors, while still offering significant data rate benefits. The achievable ergodic rate using these systems is verified to be close to the link's ergodic capacity.

show abstract

Section: B Submodular Searchmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Analysis of Laser and Detector Placement in Incoherent MIMO Multimode Fiber Systems

Appaiah¹,

Zisman²,

Das³

et al. 2014

J. Opt. Commun. Netw.

View full text Add to dashboard Cite

show abstract

“…This is because of the evaluation of the best combination of modes that provide the highest achievable rate would require checking each combination of L rows of H, thus requiring O(D L ) comparisons. To reduce this complexity, we adopt a greedy approach to select the modes, similar to the antenna selection paradigm described in [15]. Such an approach, while suboptimal, reduces the number of comparisons required to O(DL), which is much smaller.…”

Section: Sub-modularity and Mode Selection At The Receivermentioning

confidence: 99%

Low Complexity Equalization Using Mode-Subset Selection in MIMO Multimode Fiber Links

Sanila¹,

Appaiah²

2014

2015 IEEE Global Communications Conference (GLOBECOM)

View full text Add to dashboard Cite

Optical fiber links have enabled the transmission of very high data rates over both short-range and long-range links. However, the recent growth in the bandwidth demand has created the need to enhance their data rate capacities. While modern modulation techniques and multiple-input multipleoutput (MIMO) based mode-division multiplexing (MDM) techniques with multimode fibers (MMFs) have been demonstrated to be promising approaches to achieve higher data rates, the computational complexity required to perform data decoding at these high data rates makes these systems prohibitive. In this paper, we model the channel characteristics and modedependent losses of few-mode and many-mode MMFs and study the data rates obtained if only a few modes are chosen for decoding at the subset. Such an approach promises to reduce the decoding complexity while not sacrificing data rate significantly. Simulations reveal that a subset selection approach leads to large savings in computation while not compromising data rate significantly. In particular, the energy-per-bit requirement can be reduced by over 50% in few-mode fibers and upto 70% in large-core multimode fiber links, when compared to conventional MDM decoding.

show abstract

“…It was shown in [13] that C(L) is a monotone submodular in the columns of β, Theorem(1). The proof of Theorem (1) involves showing C(L) is equal to entropy of some random variable.…”

Section: Definition 2 (Monotone): a Set Functionmentioning

confidence: 99%

Mutual Information Based Output Dimensionality Reduction

Pandey

Vaze

2014

2014 IEEE International Conference on Data Mining

Self Cite

View full text Add to dashboard Cite

Given a large dimensional input and output space, even simple regression is prohibitively costly. Dimensionality reduction in the output space is important for efficient learning and prediction as modern paradigms, e.g. topic modelling, image classification, etc., have extremely large output spaces. In contrast to input dimensionality reduction, dimension reduction in output side is complicated. We propose, mutual information based output dimensionality reduction, that takes into account the relationship between the input and the output which is essential for regression and classification problems. Our method selects those labels to form the compressed label space that typically have the maximum mutual information with the input. Selecting the best subset is computationally hard, but we provide a polynomial time algorithm with provable approximation guarantee. We conduct experiments on seven multi-label classification datasets. Results show our method performs better than existing methods on some datasets.

show abstract

Sub-Modularity and Antenna Selection in MIMO Systems

Cited by 34 publications

References 27 publications

Analysis of Laser and Detector Placement in Incoherent MIMO Multimode Fiber Systems

Analysis of Laser and Detector Placement in Incoherent MIMO Multimode Fiber Systems

Low Complexity Equalization Using Mode-Subset Selection in MIMO Multimode Fiber Links

Mutual Information Based Output Dimensionality Reduction

Contact Info

Product

Resources

About