On the Throughput Maximization of Sensor Networks using Data Aggregation and Reduction Abstract—Researchers have proposed aggregation models for Wireless Sensor Nodes (WSN). These models allow the sampling of data at the sensor nodes’ internal memory. The aggregated data is transmitted to the sink node when some predefined conditions are fulfilled. It is very likely that monitored data samples may remain unchanged for significant time depending upon the environmental conditions being monitored. This paper suggests a data reduction algorithm which exploits the non- variability in the aggregated data samples at the sensor nodes. The proposed data reduction algorithm allows the accommodation of more information such as video data in WSN packets. The throughput of such kind of wireless sensor network is investigated in this paper. Results show that the data aggregation and reduction can increase the maximum throughput of WSN. Keywords-Sensor node; Slotted ALOHA; packet generation rate; data aggregation; data repetition. I. INTRODUCTION Wireless sensor networks (WSN) are playing an increasingly important role in physical parameter monitoring applications. A WSN consists of battery-powered nodes, equipped with sensors for taking readings (e.g. temperature, light, or even pictures), and a radio link for communicating with neighboring nodes over short distances. For designing a multiple access of a wireless sensor network, two factors are very important to consider. First, an efficient and distributed radio resource sharing scheme and secondly, the access scheme should be power efficient. The Slotted ALOHA is the most power efficient multiple access scheme [1]. However, its efficiency can be degraded due to excessive collisions. Many collision reduction algorithms have been proposed by many researchers. One of the effective ways of collision reduction is to use binary tree algorithm [2]. In Slotted ALOHA access scheme, reducing the number of retransmission trials is another effective technique to reduce the collisions among packets [3- 11]. Limiting number of retransmission trials reduces the aggregate traffic and thus reduces the collisions. Generally, the sensor nodes work on sample-and-transmit model. Data aggregation techniques have been proposed to improve the energy efficiency of WSN [12-14]. The aggregation technique will prevent the unnecessary continuous transmission of monitored data. Zechinelli et al proposed a stream aggregation model in which a sensor temporarily stores samples in a history located in its RAM instead of immediately transmitting the samples to the sink node [15]. The aggregated data is transmitted only in one of the following cases: when the history is full, when the size of stored data reaches the optimal packet size or the maximum packet payload, or when a query is received. It is very likely that the monitored or recorded data samples at the sensor node history may remain intact for a deterministic or un-deterministic time period depending upon the nature of the environment under surveillance. This paper presents a data reduction algorithm which exploits the non- variability of monitored data samples in a node‟s data history. The main idea is to send the difference in data samples to the sink. This may lead to pertinent impact on the energy efficiency of WSN under consideration. The paper is organized as follows. Section II describes the details of the proposed data reduction algorithm. Section III presents the impact of proposed data reduction algorithm on the throughput of slotted ALOHA based sensor networks. The conclusion is provided in Section IV. II. DATA REDUCTION ALGORITHM FOR WSN Zechinelli et al presented an energy aware data aggregation model in which temporal database is implemented in the sensors [15]. Each sensor node (SN) maintains history of n samples in SN‟s RAM. A special counter is used in each SN to estimate the size of the history. The main objective of this aggregation model is to achieve WSN‟s energy efficiency by minimizing the packet transmissions. It is very likely that the content of the temporal database maintained at each sensor node may remain intact for considerable time period depending upon the environment under surveillance. For instance, in animal surveillance and tracking system, the data captured by cameras with infrared sensor may remain constant for considerable time period. Also, the SN monitoring environmental parameters such as temperature, humidity etc. will have similar non-variability possibilities. The objective of this paper is to suggest a data reduction (DR) algorithm by exploiting the data repeatability at the sensor nodes. The proposed data reduction algorithm aims to achieve WSN energy efficiency by transmitting reduced WSN packets. A. DR Algorithm To implement the proposed algorithm, following assumptions are made: Syed Misbahuddin Electronics and Communication Department College of Engineering at Al-Lith Umm Al-Qura University, Saudi Arabia doctoyedmisbah@yahoo.com Jahinger H. Sarkar Electronics and Communication Department College of Engineering at Al-Lith Umm Al-Qura University, Saudi Arabia mjsarker@uqu.edu.sa Muhammad T. Simsim Electronics and Communication Department College of Engineering at Al-Lith Umm Al-Qura University, Saudi Arabia msimsim@uqu.edu.sa 978-1-4673-6404-1/13/$31.00 ©2013 IEEE 115