Research On Indoor Tracking And Positioning Technology Based On RSSI

In wireless sensor networks,mobile target localization is a basic function.This paper studies the indoor localization technology based on RSSI,and proposes two effective indoor tracking and localization algorithms to solve the existing problems.Firstly,a dynamic ring structure based target tracking and localization algorithm is designed to solve the problem that RSSI value is susceptible to environment and thus affects the localization accuracy.The algorithm constructs multiple ring structures within the sensing range of the sensor networks,and genetic algorithm based on elite reservation strategy is used to set the ring width to optimize the ring structure;based on this,a binary coding algorithm is proposed to determine the target occurrence area,the algorithm establishes the unique mapping between the target occurrence area and the binary coding,and uses the density-based clustering method to solve the ambiguity of the target occurrence area;in order to optimize the target localization results,this paper further proposes a two-stage centroid localization algorithm,which is especially suitable for the case where the target area is irregular.At last,the superiority and effectiveness of the proposed algorithm are verified by experiments.The effects of dynamic ring width,grid size and other parameters on the localization results are analyzed experimentally,and the localization accuracy of the target under different motion trajectories is investigated.Secondly,this paper studies RSSI fingerprint database based indoor tracking and localization technology.Firstly,the effects of three different fingerprint generation methods on the accuracy of target localization are studied.On this basis,a principal component analysis based target tracking and localization algorithm is proposed.The algorithm first collects the original RSSI values and forms them into the RSSI vector to form the original fingerprint database.On this basis,the RSSI quantization method is proposed,which uses exhaustive search and genetic algorithm based on elite preservation strategy to determine the optimal threshold,and realizes the quantization processing of fingerprint database through the threshold comparison.Next,this paper analyzes the principal component of the quantized fingerprint database,generates the reduced dimension fingerprint database,and uses it to locate the mobile target.Finally,the influence of three fingerprint generation methods on the target accuracy is analyzed through experiments,which verifies the superiority and effectiveness of the proposed algorithm.