Studies on the performance of correlation based energy efficient data compression algorithms for wireless sensor network

Abstract

Wireless Sensor Networks (WSN) is foreseen to become a wireless technology application of major importance in the future. WSNs differ from traditional wireless networks in that they are typically self organizing with a potentially huge number of randomly deployed battery driven small nodes. Failure of the sensor nodes also leads to decrease in the lifetime of the sensor network. The success of the Zigbee standards also demonstrates enormous market potential. This research work reports the simulation studies of energy efficient data compression algorithms for wireless sensor network. In this research work the energy conservation is done by developing compression algorithms that utilize the existing spatial and temporal correlation of the sensed data. Dual prediction algorithm is exploited in this research to reduce the number of transmissions by the sensor node. An Enhanced Normalized Least Mean Square (ENLMS) filter suitable for dual prediction algorithm is proposed and the hardware implementation of the algorithm is performed. The sink node during the clustering phase uses clique partitioning algorithm to cluster the nodes. During the data collection phase the sink node uses dual prediction algorithm to predict the data. This method for data transmission reduces the average energy consumption of the network. The algorithm is simulated using MATLAB wireless simulator and the results prove decrease in average energy consumption of the network for highly correlated data. Low power implementation of the Viterbi decoder using double edge triggered flip flop and clock gating technique is implemented in this research. Simulation of the proposed algorithm is performed using Xilinx software and the performance analysis shows that the modified Viterbi algorithms with low power techniques greatly reduce the power. The proposed algorithms were successfully simulated and the results indicate that exploitation of spatial and temporal correlation in sensed data is a promising technology for energy efficiency in sensor networks