Research On Collaborative Localization Technology Of Wireless Sensor Networks In Cluster Deployment

Wireless sensor network is an ad-hoc network formed by a large number of nodes with different sensing capabilities in a multi-hop transmission mode.Localization technology is of vital importance in the research of wireless sensor networks.Due to different positioning scenarios,localization technology requirements are also different.For the lunar surface with complex terrain and harsh environment,the data collected by the sensor nodes at different locations varies greatly,so the data information collected by each node needs a more accurate geographic location.For this special positioning scenario,not only a large monitoring range is required,but also there are some other problems and challenges: due to the limitation of the load,the number of deployed nodes is limited result in a small number of neighbor nodes of the unknown node;the cost of anchor nodes is high,considering the overall cost of the network,the number of anchor nodes is relatively small;the number of neighbor anchor nodes even neighbor nodes of some nodes on the edge of the network is less than three,which cannot complete the positioning.Therefore,a large-scale monitoring scheme based on cluster deployment is used in this paper,and an effective localization algorithm is also required to meet the localization requirements.Aiming at the above problems,this paper proposes a collaborative localization algorithm RCCL based on RSSI and colinearity and furthermore an improved RCCL algorithm based on confidence.By designing simulation experiments,this paper demonstrates the effectiveness of the proposed algorithm.The specific research contents of this paper are as follows:First,a collaborative localization algorithm RCCL based on RSSI and colinearity is proposed.This paper is designed to learn the lunar surface signal fading model by the interactive information within the initial anchor nodes and then obtains the ranging model.Through theoretical analysis,the reason for the large localization error is explained when the anchor node coordinates assisting positioning are colinear,demonstrating the disadvantages of maximum likelihood.And in view of this situation,a colinear localization method based on the point intersection between “communication circle” is proposed.In addition,due to the low density of anchor nodes and the small number of nodes,some unknown nodes that cannot complete the positioning adopt a new virtual neighbor anchor node localization method to improve the node localization rate.The RCCL localization algorithm is compared with the maximum likelihood estimation method based on RSSI,as well as the improved triangular centroid algorithm based on RSSI and colinearity through experimental simulations,which shows that the algorithm proposed in this paper has a large localization accuracy and node localization rate.The improvement proves the effectiveness of the algorithm.Then analyzing the error sources of the RCCL localization algorithm,this paper makes improvements from three aspects: RSSI ranging,anchor node selection,and iterative refinement.First,the method of Kalman filtering and mean filtering is used to preprocess the RSSI measurement value;then a confidence formula based on reverse relative localization error is proposed in accordance with that the unknown nodes can select the node that have completed the positioning to upgrade.The final estimated coordinates of the unknown nodes are obtained in a method of iterative manner.Experimental simulation shows that the improved RCCL algorithm based on confidence improves the localization performance by about 27% compared with the RCCL algorithm.