Research On WLAN Indoor Localization Algorithm Based On FCMM And PCA

The rapid development of wireless Local Area Network and the wide application of intelligent mobile terminals provide a broader space for the development of indoor localization technology.WLAN fingerprint location system based on Received Signal Strength of Wireless Access Point is developing rapidly.At the same time,it faces the problems of decreasing location accuracy and increasing location complexity.Aiming at the problem of increasing complexity of location matching caused by the increase of indoor AP,this paper uses K-means clustering to reduce the search space of fingerprint,and establishes a location model in each independent location sub-region.Aiming at the problem that K-means clustering algorithm is easy to fall into local optimum solution and affects the accuracy of location,this paper improves K-means clustering algorithm.A new clustering algorithm of firefly based on Chaos theory and Max-min distance is proposed for fingerprint location.The Chaotic variables with regularity and ergodicity and firefly algorithm with stochastic optimization performance are introduced in this algorithm.The selection of initial clustering centers and the clustering process of sample points are optimized respectively by chaotic search and intelligent location updating.The simulation results show that FCMM algorithm is more effective than other algorithms in solving local optimal solution problems.Aiming at the problem that location fingerprints are difficult to match because of the large fluctuation of AP,and the strong correlation of RSS signals from adjacent AP results in redundant information,which increases the location complexity,an algorithm for joint stable AP optimization and PCA feature extraction is proposed in this paper.In the off-line data acquisition stage,this algorithm filters the volatile AP by stable AP optimization method,then divides the high-stability AP into different locating sub-regions by FCMM algorithm,and extracts the locating feature of RSS information in each locating sub-region by principal component analysis,so as to reduce the correlation of fingerprint signals and remove redundant information.Comparing with other fingerprint location methods,the experimental results show that the algorithm combined FCMM and PCA proposed in this paper can effectively reduce the location time and location error,reduce the complexity and improve the accuracy.