Dynamic sensing optimization strategy for mobile nodes deployment in wireless sensor networks
Sheng Wang,Xue Wang,Junjie Ma
DOI: https://doi.org/10.1117/12.717495
2006-01-01
Abstract:Sensor nodes deployment problem is one of the fundamental issues in wireless sensor networks (WSNs) which should consider a tradeoff among several metrics, such as coverage area, reliability, accuracy, lifetime etc. The mobile sensor nodes which can relocate themselves can be used to optimize the nodes deployment under various kinds of situations. Because coverage area is hard to be calculated by analytical method, an areas division method is introduced to evaluate the coverage area metric for simplifying calculation. Then we introduce a practically feasible combined metric which refers to coverage area, reliability, accuracy and lifetime, which uses areas division, detecting reliability, Mahalanobis distance and energy entropy as metric functions. Here, nodes deployment is considered as an optimization problem. Particle swarm optimization (PSO) algorithm, which has a series of advantages, such as, high-speed regional convergence, efficient global searching ability, and so on, is suitable for solving multi-dimension function optimization in continuous space. So we adopt PSO for nodes deployment optimization where the combined metric is considered as fitness function. Because the combined metric is multiform and changeable in PSO, we can adopt different combined metrics for different applications, while other strategies just consider the coverage area in nodes deployment. The experimental results verify that the PSO based mobile nodes deployment strategy has good performance in quickness, which can improve the capabilities of WSNs and dynamically adjust the deployment according to the changes of situation, especially when some areas need multiple-node-measurement.