The Research On WSN Indoor Location Based On Access Points Selection

With the development of the indoor localization of Wireless Sensor Network,the demand based on location service is becoming more and more widely,accordingly it drives a great attention of scientific research,business and industrial departments.The commercial application of Wireless Sensor Network makes it develop forward rapidly,so that it is integrated into various fields.Scientists predict that WSN will cause an information revolution and set off a new wave of industry.Nowadays,more and more researchers are studying the localization of Wireless Sensor Networks,and the indoor localization method based on WSN is becoming more and more mature.In the future,the WSN is everywhere,and it will greatly affect human lifestyle.In the indoor localization algorithm of Wireless Sensor Network,the fingerprint matching and the ranging model which based on RSSI indoor localization algorithm,have stronger application advantages.But according to the requirements of indoor environment complexity,special precision,and those uncertain factors of signal which includes refringence,diffraction,multipath propagation and so on in its transmission in air,how to optimize and improve the indoor localization based on RSSI has been a popular hot spot about the sensor indoor localization at present.In view of the existing fingerprint simplified method is mostly to analysis the original discrete RSS,or directly use RSS information such as the mean and the standard,there are discrete data represent only the signal intensity of the sampling times,lead to part of the problem of information loss,by maximizing curve similarity mechanism,put forward improved gaussian filter to the original RSS data preprocessing,by using the gaussian model matching algorithm pending a final node location.The effectiveness of the proposed algorithm is verified by the experimental structure.Aiming at solving the problem that access points of wireless indoor location have correlation,this article through the analysis of the existing fingerprint simplified algorithm,a novel improved algorithm called CMI(Continuous Mutual Information)is proposed.Considering the correlation between the access points,an access point selection algorithm based on mutual information and entropy is proposed in online phase.The correlation between access points is measured through continuous mutual information;at the same time,combining the information entropy to search the access points which contain most valid information about the location.Experiment results show that,comparing with traditional indoor positioning methods,such as WKNN and MLE,the proposed CMI algorithm has better localization accuracy.The algorithm in this paper is mainly divided into following steps:Firstly,We use the field of improved gaussian fitting method will offline stage collected at the reference point of the original discrete RSS value preprocessing,reduce the noise influence on positioning accuracy,and realize the continuous and discrete signal the intrinsic relationship between mining data;Put forward the Priority degree function is used to examine the AP regional features;Secondly,on the online phase,Implements a gaussian fitting method for the original RSS.The correlation between access points is measured through continuous source joint mutual information;at the same time,combining the information entropy to search the access points which contain most valid information about the location;Finally,matching the pending node AP subset and reference point AP subset of fingerprint database,select reference points where effective AP number not less than three.Kullback-Leibler divergence is used to match position,and estimate the node position.Experimental design use the deployment of AP TL-WA501+,selects ASUS series X8AIN Notebook PC as RSS receiving device,use the author designed to receive the operating system running on Windows 7 and carry out the simulation experiment by the mobile phone collection software on the mobile phone terminal of the Android system.Through the realization of data analysis,the algorithm not only have a greater improvement in the positioning accuracy,but also due to the indoor scene factors considered more comprehensive it also has a certain extension.