Investigation of Energy Cost of Data Compression Algorithms in WSN for IoT Applications

Abstract

The exponential growth in remote sensing, coupled with advancements in integrated circuits (IC) design and fabrication technology for communication, has prompted the progress of Wireless Sensor Networks (WSN). WSN comprises of sensor nodes and hubs fit for detecting, processing, and communicating remotely. Sensor nodes have limited resources such as memory, energy and computation capabilities restricting their ability to process large volume of data that is generated. Compressing the data before transmission will help alleviate the problem. Many data compression methods have been proposed but mainly for image processing and a vast majority of them are not pertinent on sensor nodes because of memory impediment, energy utilization and handling speed. To overcome this issue, authors in this research have chosen Run Length Encoding (RLE) and Adaptive Huffman Encoding (AHE) data compression techniques as they can be executed on sensor nodes. Both RLE and AHE are capable of balancing compression ratio and energy utilization. In this paper, a hybrid method comprising RLE and AHE, named as H-RLEAHE, is proposed and further investigated for sensor nodes. In order to verify the efficacy of the data compression algorithms, simulations were run, and the results compared with the compression techniques employing RLE, AHE, H-RLEAHE, and without the use of any compression approach for five distinct scenarios. The results demonstrate the RLE’s efficiency, as it surpasses alternative data compression methods in terms of energy efficiency, network speed, packet delivery rate, and residual energy throughout all iterations.

LPRLC: Linear Predictive Run Length Coding to Improve Energy Consumption of WBANs

“…The third group of compression is a hybrid that is made from the combination of lossless-lossless [32][20], lossy-lossy, and from both groups lossless and lossy [21][25] [29][34][50]. Today, many groups of hybrid methods are proposed to benefit from the advantages of both methods in reducing data volume.…”

“…Today, many groups of hybrid methods are proposed to benefit from the advantages of both methods in reducing data volume. The H-RLEAHE [32] algorithm is one of the lossless-lossless hybrid methods. It has two phases, the first phase of data encoding through RLE and the second phase of tree construction using Huffman in the sender and receiver, both of those use the same initial node and show the nodes that are Not Yet Sent (NYT) with zero weight.…”

“…The data is collected in cluster head or sink and data labeling happens after recovery using a backpropagation neural network. In H-RLEAHE [32], there is a combination of RLE and Huffman algorithms in different scenario, both of which are entropy-based. The RLE algorithm alone has a higher compression rate and outperforms than H-RLEAHE.…”

“…In this regard to solve the problem of unimportant and repetitive data and high energy consumption in the sensors layer, a hybrid method is presented [31] [32]. LPRLE has three phases: prediction, error calculation, and error compression.…”

“…LPRLE is compared with different algorithms of lossless coding: Huffman [8], LZW [17], and Arithmetic [13]. Results show the LPRLE with LPC and RLE has the greatest energy saving on average of 98% by keeping the proper compression ratio in five biosensors than DDCA [20], SZ with Huffman [18] and H-RLEAHE [32].…”