ALBERT

Description: Sensors, Vol. 18, Pages 2031: A Network Equivalent-Based Algorithm for Adaptive Parameter Tuning in 802.15.4 WSNs Sensors doi: 10.3390/s18072031 Authors: Yipeng Wang Wei Yang Ruisong Han Kaiming You Previous studies have shown that in many wireless sensor network applications the IEEE 802.15.4 carrier sense multiple access with collision avoidance (CSMA/CA) mechanism with default parameters cannot guarantee the constraints of reliability, time efficiency, or energy efficiency. Although many adaptive parameter tuning algorithms have been proposed, many of them cannot correctly identify the changes of the network condition and are unable to effectively perform the parameter tuning operation. Considering the randomness that CSMA/CA brings about, for most of the proposed algorithms, it is a challenge to distinguish significant violations that were caused by actual changes of the network from the general fluctuations that were due to CSMA/CA. In this paper, we propose a lightweight algorithm called the network equivalent adaptive parameter tuning (NEAPT) algorithm. It is fully distributed and can work without any predefined information or acknowledgement. NEAPT not only takes reliability as an evaluation of a network condition, but it proposes a synthetic value, called the equivalent node number, and takes it as another reference for a network condition. Simulation results show that by taking both reliability and the equivalent node number into consideration, NEAPT can effectively identify the network changes and provide adequate and steady performances for wireless sensor networks (WSNs) in both stationary and dynamic conditions.

Description: Clustering is an effective technique used to reduce energy consumption and extend the lifetime of wireless sensor network (WSN). The characteristic of energy heterogeneity of WSNs should be considered when designing clustering protocols. We propose and evaluate a novel distributed energy-efficient clustering protocol called DCE for heterogeneous wireless sensor networks, based on a Double-phase Cluster-head Election scheme. In DCE, the procedure of cluster head election is divided into two phases. In the first phase, tentative cluster heads are elected with the probabilities which are decided by the relative levels of initial and residual energy. Then, in the second phase, the tentative cluster heads are replaced by their cluster members to form the final set of cluster heads if any member in their cluster has more residual energy. Employing two phases for cluster-head election ensures that the nodes with more energy have a higher chance to be cluster heads. Energy consumption is well-distributed in the proposed protocol, and the simulation results show that DCE achieves longer stability periods than other typical clustering protocols in heterogeneous scenarios.