Compress sensing algorithm for estimation of signals in sensor networks
- PDF / 2,958,281 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 90 Downloads / 212 Views
(0123456789().,-volV)(0123456789(). ,- volV)
Compress sensing algorithm for estimation of signals in sensor networks Juan Martinez1 • Jose Mejia1 • Boris Mederos1 Jose´ Antonio Marmolejo-Saucedo3
•
Alberto Ochoa1
•
Oliverio Cruz-Mejı´a2
•
Ó Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract In this research, we present a data recovery scheme for wireless sensor networks. In some sensor networks, each node must be able to recover the complete information of the network, which leads to the problem of the high cost of energy in communication and storage of information. We proposed a modified gossip algorithm for acquire distributed measurements and communicate the information across all nodes of the network using compressive sampling and Gossip algorithms to compact the data to be stored and transmitted through a network. The experimental results on synthetic data show that the proposed method reconstruct better the signal and in less iterations than with a similar method using a thresholding algorithm. Keywords Sensor networks Compressive sampling Gossip algorithms
1 Introduction Sensor networks, both wired and wireless, have found a wide variety of applications, which has caused their growth, and with this, a greater amount of information to propagate [23]. Thus, there exists a constantly search for optimization in both speed and in the amount of data to be transmitted. In a network with thousands of nodes, for unify their information it would be necessary at least n transmissions, which n is the number of nodes, in practice this could be very slow or not at all useful, also this can carry out other problems such as: complications to detect possible failures, maintenance, or even to detect possible attacks on the network [12]. To solve this problems, an alternative is the use of compressed sensing Electronic supplementary material The online version of this article (https://doi.org/10.1007/s11276-019-02031-5) contains supplementary material, which is available to authorized users. & Oliverio Cruz-Mejı´a [email protected] 1
Universidad Auto´noma de Ciudad Jua´rez, Ciudad Jua´rez, Mexico
2
Universidad Auto´noma del Estado de Me´xico, Toluca, Mexico
3
Facultad de Ingenierı´a, Universidad Panamericana, Augusto Rodin 498, 03920 Ciudad de Me´xico, Mexico
(CS) algorithms [4, 10], which seek to reconstruct a signal starting from a much smaller amount of data, this is possible by expressing the signal in sparse domain (Fourier, Wavelet, etc.), where most of its elements are null, this could be beneficial in both speed, amount of data to be transmitted, power savings and facilitate the analysis of information. The use of optimization and CS has been used in a distributed setting in several works. In [21] it is presented a distributed projected consensus algorithm were nodes combine their local average with projection on their individual constraint sets. In [19] are proposed several strategies based on distributed iterated hard thresholding algorithm over a network that employ diffusi
Data Loading...
