With the development of wireless technology and micro-sensor technology, wireless sensor network becomes very important in information collection and treatment. Node localization is the most critical supporting technologies of applications in wireless sensor networks. As wireless sensor network is often used in harsh environment or people can not reach the area, so it is not possible by manual deployment to determine the location of the node, however, node localization that is implemented by the use of GPS or other external device is limited by the cost and energy consumption of nodes. It can only be by a few nodes equipped with GPS location device, then according to their communication with other nodes, the location of the unknown node is completed by localization algorithm.In wireless sensor network, node localization algorithms are divided into range-based localization algorithm and range-free localization algorithm. Range-based localization algorithm directly uses hardware ranging device to measure the distance between the nodes, then gets the coordinates of the unknown node by calculating the relative location between the unknown node and the anchor node. The category of above algorithms is simple and high localization accuracy, but the ranging device will increase the cost of the sensor node, ranging process will increase the node energy consumption, therefore, lifetime of wireless sensor network is affected. Range-free localization algorithm calculates the unknown coordinates of the node by the communication of nodes and distance between nodes. Among a number of range-free localization algorithms, DV-Hop algorithm is one of the most concerned-about algorithms.In DV-Hop algorithm, estimated distance is calculated by the hops between the unknown node and the anchor node multiplying the mean distance per hop. Distance estimation error is affected by network topology and density of nodes in the network, so the lower density, the higher localization error. In order to solve these problems, in this paper, combining DV-Hop and Particle Swarm Optimization, a simplified particle swarm algorithm based on the variation of the degree of aggregation is proposed to optimize DV-Hop algorithm, utilizing the feature of iterative optimization in Particle Swarm Optimization, restraining the effect of distance estimation error accumulation on localization accuracy. tsPSO eliminates the divergence problem caused by particles due to speed entry by simplifying the standard particle swarm algorithm speed formula, so the slow convergence rate and the low accuracy is solved. Meanwhile, in order to avoid premature convergence in algorithm optimization process, the extreme value is random turbulent by setting the threshold, so that particle optimization is guided.Unstable convergence phenomenon in algorithm optimization process is caused by tsPSO that uses extreme turbulence. In order to overcome the drawback, a simplified particle swarm algorithm based on the variation of the degree of aggregation is presented in this paper, firstly, the degree of aggregation of particle swarm is calculated, then the particle variation is selected by combining with the similarity between the particle and the diversity of the population is increased, so the premature convergence of the algorithm is avoided. Finally, the DV-Hop algorithm is optimized by the new algorithm instead of tsPSO algorithm.Experimental analysis is done by Matlab simulation software. Experimental results with standard DV-Hop localization algorithm and based on the standard PSO-DV-Hop algorithm are compared. It indicates that node localization accuracy is significantly improved by no matter tsPSO algorithm or csPSO algorithm, the effect is more obvious especially when a very low percentage of anchor nodes. Compared with DV-Hop localization algorithm based on standard PSO, in this paper, the localization algorithm on the convergence rate has an obvious advantage. csPSO algorithm has better performance than tsPSO algorithm in the term of convergence stability. |