Affiliations: [a] School of Artificial Intelligence, Hebei University of Technology, Tianjin | [b]
Xingtai Polytechnic College, Xingtai, Hebei, China
Correspondence:
[*]
Corresponding author. Taihang Du, School of Artificial Intelligence, Hebei University of Technology, Tianjin 300130. E-mail: thdu_tj@163.com.
Abstract: In view of the high price of receivers in indoor radio interference sources, a double fingerprint database acquisition and location system based on wireless sensor networks is designed according to the design aim of low input and high efficiency. To meet the demand of the number of at least AP positioning in the positioning area, the main library collection in the overall positioning of fingerprint point and signal intensity, using the improved fuzzy C clustering (Fuzzy C Means, FCM) algorithm and support vector machine (Sport Vector Machine, SVM) location of interference source location and area, area after the parade AP moving to the main fingerprint lock, before relying on offline collected local fingerprint information database, according to the Euclidean distance model of the objective function, the positioning process into local optimization and application of genetic algorithm (Genetic Algorithms GA) and Particle Swarm (Particle Swarm Optimization, PSO GA-PSO) algorithm to calculate model the optimal solution, converting the location coordinates of the interference source. The experimental results show that the confidence level of the non-sight distance positioning accuracy of 1.3m obtained at least AP is 76.3%, and the method used in the study is of high practical value.
Keywords: Indoor positioning, clustering method, simulated annealing simplex fusion
algorithm, GA-PSO fusion algorithm, cruise AP signal receiver
