Design And Implementation Of Low Power Consumption Indoor Localization System Based On RSSI

With the development of such satellite positioning systems as GPS and Beidou,outdoor localization technology has been widely used in daily life.Similarly,indoor localization,as an important part of localization,has a broad application prospect in many fields of people's work and life.However,on the one hand,due to the obstructions from buildings,the satellite signals for indoor localization could not be well received.On the other hand,in the complex indoor environment,there are signals reflection,diffraction and multipath effects,which lead to the fact that so far,the application of indoor localization technology has not yet been popularized.On the contrary,with the development and progress of society,the demand for mature indoor localization technology is becoming more and more urgent.Based on the fingerprint localization technology,a set of indoor localization system is studied and designed in this paper,the realization of which could be divided into two phases,namely,the off-line collection phase and the online localization phase.The first phase is mainly used to collect RSSI(Received Signal Strength Indication)at each reference position to build fingerprints and then to store them in the fingerprint database.Since the directly collected RSSI contains a lot of noise because of some interfering factors in the indoor environment,the RSSI is needed to be pre-processed before stored in the fingerprint database.Accordingly,the filtering effects of different widely used filtering methods are firstly discussed and analyzed through experiments,finding out that Calman filter has a better filtering effect on random noise.On the basis of the experiments and analysis,and the consideration of the various noise in the original RSSI,a set of filtering steps is put forward in this paper where the residual filter,Kalman filter and mean filter are successively adopted.The main task of the online localization phase is to match the real-time fingerprint with the fingerprint in the fingerprint database by using appropriate matching algorithm,thus estimating the target localization according to the matching results.Specifically,in this paper,the shortcomings of the Nearest Neighbor algorithm are improved and an adaptive Nearest Neighbor matching algorithm based on the distance of transmitting nodes is proposed.The original matching algorithm and the improved matching algorithm are compared through experiments,the results of which indicate that compared with the original matching algorithm,the localization accuracy is improved by about 5%-10%,and thematching efficiency is improved as well in the improved matching algorithm.Finally,based on the ZigBee wireless network,and combining the filtering steps with the improved matching algorithm proposed in this paper,the fingerprint-based indoor localization technology is implemented,the localization accuracy of which in the experimental environment is analyzed through the experimental results.