讨论组:无线传感器网络组
标题:Major Coefficients Recovery: a Distributedly Compressed Data Gathering Scheme fo
演讲人: 许力文,张晨
时间: 2011-03-22 10:30-2011-03-22 12:00
地点:FIT 1-222

内容:

Title: Major Coefficients Recovery: a Distributedly Compressed Data Gathering Scheme for Wireless Sensor Network

Abstract: or large-scale sensor networks deployed for data gathering, energy efficiency is critically required. Elimination the data correlation is a promising technique for energy efficiency. Compressive Data Gathering (CDG) which uses linear online encoding to compress data correlation is a breakthrough in this area. However, the CDG scheme uses a uniform pattern in data transmission, where all nodes transmit the same amount of data regardless of their hop distances to the sink. This pattern has transmission overhead $O(kn\log{n})$ in the 2-D networks. In this paper, a major coefficient recovery (MCR) scheme is proposed, where Discrete Cosine Transformation (DCT) is applied distributedly to the original data sensed. A non-uniform data transmission pattern is proposed by exploiting the energy concentration property of DCT and QR decomposition techniques, so that sensors with larger hop-count can transmit less messages for network energy efficiency. The sink node recovers only the major coefficients of the DCT to recover the environment observantion accurately. Because sensors with larger hop-count are much more than the sensors with smaller hop-counts in the data collection networks, MCR reduces the transmission overhead to $O(kn-k^2)$, which is $O(logn)$ better than CDG. The recovery performances of MCR are verified by extensive simulations.

 

Title: Compressed Sensing of Gauss-Markov Random Fields with Wireless Sensor Networks

 

Abstract: We propose a scalable and energy efficient method for reconstructing a ‘sparse’ Gauss-Markov random field that is observed by an array of sensors and described over wireless channels to a fusion center. The encoder is universal, i.e.invariant to the statistical model of the source and the channel, and is based on compressed sensing. The reconstruction algorithms exploit the a-priori statistical information about the field and the channel at the fusion center to yield a performance comparable to information theoretic bounds.