Investigation of Parameter Effects on Virtual-Spring-Force Algorithm for Wireless-Sensor-Network Applications

Abstract

Virtual-force algorithms (VFAs) have been widely studied for accurate node deployment in wireless-sensor-network (WSN) applications. Their main purpose is to achieve the maximum coverage area with the minimum number of sensor nodes in the target area. Recently, we reported a new VFA based on virtual spring force (VFA-SF) and discussed in detail the corresponding efficiency via statistical analysis. The optimized strategy by adding an external central force (VFA-SF-OPT) was presented, which effectively eliminates the coverage hole or twisted structure in the final network distribution. In this paper, the parameter effects on VFA-SF and the VFA-SF-OPT were further investigated: (1) Node velocity dramatically affects the convergence rate of the node-deployment process. (2) A suitable external central force improves equilibrium distance and reduces energy consumption. (3) The effects of VFA-SF and VFA-SF-OPT for different types of obstacles are discussed. Generally, by choosing suitable parameters, both VFA-SF and VFA-SF-OPT can effectively improve node deployment and energy consumption for the whole sensor network. The results give important insight in parameter selection and information fusion in the application of a large-scale WSN.

Node Deployment Optimization for Wireless Sensor Networks Based on Virtual Force-Directed Particle Swarm Optimization Algorithm and Evidence Theory

“…Theoretical analysis has shown that a hexagonal structure is the best two-dimensional network for providing the greatest coverage area with the fewest sensor nodes and the least amount of energy usage [ 21 ]. Yu et al [ 22 ] reported a novel VF model based on virtual spring force (VFA-SF) and examined the associated efficiency in detail using statistical analysis. To place the self-consistent nodes in large-scale WSNs, a hybrid optimization algorithm based on two separate VF algorithms inspired by the interactions among physical sensor nodes was presented in [ 23 ].…”