Compressed sensing at the ICC 2010

After a short vacation today I'd like to finish the review of this year's ICC. I was curious if the trend on compressed sensing I had seen at the ICASSP continued at the ICC... it seems that the ratio of papers related to compressed sensing in communications is smaller than in other fields yet. Here some of these papers:

Husheng Li presents in Reconstructing Spectrum Occupancies for Wideband Cognitive Radio Networks: A Matrix Completion via Belief Propagation a distributed spectrum monitoring scheme based on belief propagation. While this paper is not directly related to compressed sensing, other matrix completion techniques use compressed sensing theory by minimizing the nuclear norm of the matrix of interest.

In Distributed Compressive Spectrum Sensing in Cooperative Multi-hop Cognitive Networks, Z. Fanzi et al. present an elaborated cognitive radio multihop network model where the adquisition at the individual nodes is performed by means of compressed sampling. They divide the exchanged information into a common support and a series of innovations seen at individual nodes.

Also related with compressed sensing is the work Space-Time Turbo Bayesian Compressed Sensing for UWB Systems by D. Yang et al. They propose an algorithm for the joint reconstruction of ultra-wideband (UWB) signals based on the sparsity derived from both spatial and temporal redundancies.

More theoretical is the work RIP-fulfilling Complex-Valued Matrices by A. Amini et al. The abstract reads:

Although the theoretical results in the field of compressed sensing show that large classes of random matrices fulfill the so called Restricted Isometry Property (RIP) with high probability, only a few deterministic matrix designs are known. In this paper, we generalize one of the recent schemes based on binary BCH codes to p-ary codes which are useful for construction of complex sampling matrices. Though the design approach is similar, due to the use of p-ary codes (with p a prime power) and then complex matrices, the results are not similar. The new matrices are of the size (p^a - 1) × p^b using a prime power p; the previous BCH structures are the special cases for p = 2^1 which means that the new matrices provide more options in the number of samples.

In Does Compressed Sensing Improve the Throughput of Wireless Sensor Networks? Jun Luo et al. disscuss how much (in terms of throughput) can be gained by applying compressed sensing schemes at the network layer. However the results seem to be model dependent.

Linda M. Davis et al. present in Multi-antenna Downlink Broadcast using Compressed-Sensed Medium Access a communications scheme where the channel state information adquisition and user selection are performed by means of compressed sensing. Somehow related is the work Compressive Sensing for Reducing Feedback in MIMO Broadcast Channels by Syed T. Qaseem et al. The scheme prtesented is based in that only a limited number of mobile users present a channel quality above a threshold in a given time instant and spatial direction, thus the response vector is sparse in the "user domain".

A different application for compressed sensing algorithms can be found in Cooperative Sensing and Compression in Vehicular Sensor Networks for Urban Monitoring by Xiaoxiao Yu et al. In this work the authors propose an urban environment surveillance scheme utilizing vehicle-based sensors. The possible information loss due to the dynamic and unpredictable network topology is attenuated through a cooperative data sensing based on sparse random projections and a compressed sensing based recostruction.