Research On Fuzzy Theory Based Localization Algorithms In Wireless Sensor Networks

Wireless sensor network localization has been widely used in intelligent services.However,traditional localization algorithms usually require complex calculations and deserve low accuracy,and positioning error fluctuates drastically while affected by interference in the environment.This paper designs a fingerprint localization method based on fuzzy theory.After establishing a fingerprint database,the received signal strength information of the unknown node is obscured.Through data processing and information fusion,the location of the unknown node in wireless sensor networks is obtained.To improve the positioning accuracy and flexibility in different environments,a positioning algorithm based on fuzzy and DS evidence theory is proposed.The concept of membership in fuzzy theory is introduced to process the received signal strength information and analyze the distance relationship between the unknown node,reference points and anchor nodes.Adopting DS evidence theory and calculating the confidence interval.Different anchor nodes cognize the position relationship between the reference point and the unknown node in terms of trust,doubt and uncertain.Finally,evidence is synthesized to obtain the similarity between the unknown node and reference points,and then the coordinates of the unknown node is predicted.The algorithm uses the DS evidence theory model to achieve multi-dimensional information fusion.Simulation experiments show that the algorithm has small positioning error and stable performance under environmental changes.In the processing and fusion of massive data,wireless sensor network localization algorithms often sacrifice computational complexity and cost for the improvement of positioning accuracy.To handle this situation,a multidimensional Gaussian positioning algorithm based on fuzzy is proposed.According to the fingerprint database obtained during the offline phase,the fuzzy membership function is obtained,thus the fuzzy attribute matrix between the reference point and the unknown is established.The ideal point close to the unknown node is constructed,and the location deviation between the ideal point and reference are quantified based on the multidimensional Gaussian model,then the adjustment function is introduced to modify its influence on the position relationship between the reference point and the unknown node.The position similarity between the reference point and the unknown node is estimated to obtain the positioning coordinates.The algorithm operates simply,and the experimental results show that the accuracy of the positioning results is high and the range of error fluctuation is small.