A Node Localization Algorithm Based On Optimal Distance Estimation And Particle Swarm Optimization For Wireless Sensor Networks

In many wireless sensor networks applications,it is essential for the sensor nodes to automatically discover their locations.An implicit assumption underlying the Non Range-based approach is that in practice,sensor nodes have limited resource,and they are often placed with a large scale and a certain degree of uniformity.This assumption brings great challenges to localization algorithm,which seeks to strike a balance between target reliability and high efficiency in anisotropic networks with concave regions.In this dissertation,a localization scheme that works for 2D/3D networks with concave regions is presented,which is based upon Non Range-based approach,thus extending the applications to other sensor networks.The main contents of this dissertation are as follows:1)The research status of wireless sensor network localization algorithm and routing protocols are introduced.The localization technologies and their features are analyzed.Assuming sensors are placed in a 2D/3D space,possibly with holes or concave regions,the application restrictions of existing Non Range-based localization algorithm are discussed.2)A correlated-based coverage priority and energy balance probabilistic flooding algorithm(CCEP)is presented.CCEP assigns priorities for neighbors based on their relative coverage and residual energy.It also exploits the link correlation that nodes with high correlation to a common sender are assigned to a single ACK.By increasing the retransmission nodes one by one and collecting received ACKs,the expected reality and the real-time reliability are dynamically calculated and compared.After that,the minimal retransmission subsets and retransmission number are estimated.The simulation results reveal that CCEP achieves the target reliability with higher node efficiency,and reduces network communication cost.CCEP simultaneously reduces network variance of residual energy,thus prolonging the network lifetime.3)An optimal distance estimation algorithm(ODESPC)is proposed to estimate The Shortest Path Confidence.The Shortest Path Confidence is calculated based on the network connectivity and special nodes,which are on The Shortest Path Tree that is generated by the network edge nodes' flooding.ODESPC is able to improve the accuracy of distance estimation between sensor nodes.Simulations show that ODESPC works for networks with concave regions.It also improves the localization coverage ratio.4)A PSO localization algorithm based on optimal distance estimation(PSO-LAODE)for wireless sensor networks is presented.During the localization process,it corrects the average distance between nodes by using weighted average method,and optimizes the results of coordinate calculation results by using the improved Particle Swarm Optimization Algorithm.Simulation is carried out to analyze the influence on localization accuracy with a different number of initial beacon nodes.The PSO-LAODE is compared with the typical DV-Hop node localization algorithm.Simulations show that PSO-LAODE algorithm suits well to the large-scale 2D/3D sensor networks,possibly containing holes or concave regions,even with few initial beacon nodes.