Aggregation-Scheduling Based Mechanism for Energy-Efficient Multivariate Sensor Networks

• Published in: IEEE sensors journal Vol. 22; no. 16; pp. 16662 - 16672
• Main Authors: Ibrahim, Marwa, Harb, Hassan, Nasser, Abbass, Mansour, Ali, Osswald, Christophe
• Format: Journal Article
• Language: English
• Published: New York IEEE 15.08.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE); Institute of Electrical and Electronics Engineers
• Subjects: Agglomeration; Algorithms; Clustering algorithms; Clusters; Coloring-map algorithm; Data aggregation; Data collection; data transmission minimizing; Decision making; Energy conservation; Energy efficiency; Engineering Sciences; Job shop scheduling; Multivariate analysis; multivariate sensor networks; Nodes; Object recognition; Scheduling; scheduling strategy; Sensors; Signal and Image processing; Switches; Sensor networks; Job shop scheduling; Scheduling algorithms; Clustering algorithms
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2022.3189431

Recently, the world has witnessed a technology revolution in many sectors and fields with the aim to enhance the quality of life and the human security. Particularly, the integration of small sensing devices into a wide range of objects has enabled the deployment of new services and applications that made our lives smarter and safety. With such sensing-based technology, we can adequately monitor, access harsh environments, understand various phenomena and make a decision accordingly. However, sensing-based applications face several challenges beginning from the data collection at the sensors themselves until the decision-making process at the end-user. In this study, we focus on the energy conservation as major challenge in sensor applications due to the limited sensor resources that are not always possible to be replaced or recharged. We propose an aggregation-scheduling based mechanism called as AGING for Energy-Efficient multivariate sensor networks (MSN). Mainly, AGING is based on the cluster network scheme and consists into two phases: aggregation and scheduling. The aggregation phase is applied at each node and aims to reduce the huge amount of data sent periodically to the cluster-head (CH) based on a user-defined score table and a multi-aggregation mechanism. The second phase is applied at CH and uses a new scheduling strategy to switch nodes generating similar data into sleep/active modes; the CH first converts the correlated nodes into a graph before applying a coloring-map algorithm and a scheduling strategy to select the set of active nodes in next periods. Through simulations on real sensor data, we show the relevance of our mechanism in terms of saving the node energies and prolong the network lifetime compared to other existing techniques.