The Design And Implement Of Position Algorithm Based On Particle Swarm Filter In WiFi Network

What stimulate the requirement of people to the service based on location is the development of digital media business and the widespread of smart-device.The key for location based services is acquiring the position accurately and fast,therefore,which create a flourishing day for position,especially for indoors position.At the same time,WiFi is becoming a widely applicable technology of wireless network for the advantage of wireless and flexible topological structure,which makes it ubiquitous.Therefore,more and more researchers provide a location system for people under the consideration of combining WiFi and the location technology.However,we often get a terrible location result or can't location at all when use the outdoors location method to position indoors directly.This thesis puts forward one accurate position method which suit for indoors location.For more detail,we achieve accurate position indoors based on the dynamical adaption and the extreme value variation particle swarm filter algorithm.The accurate position algorithm has been put forward in this thesis has two mechanism,rough location and accurate location.We select scenario analysis to realize the rough location system in this thesis,which includes two steps,Firstly,package the coordinate of a place and the information of the signal available in the position to get the fingerprint and save it into database.Secondly,compare the real-time fingerprint with those in the fingerprint database,and find the one as the target position who has the least distance with the real-time one,namely roughly location.Based on rough mechanism,accurate location mechanism uses the dynamical adaption and the extreme value variation particle swarm filter algorithm to locate target in preciseness.Put forward several improvements for location process under the consideration of some possible problems when realizing and application,just as the follow:? Reduce the terminal equipments' influence to the fingerprint in difference method after making a good knowledge of the consistent of fingerprint data and its transport path.? Advance quality of the difference signal into fingerprint,by employing wavelet filter to remove noise and exception.? Change the method of selecting initial center of k-Means from randomly to a well-distributed way in the fingerprint database,to raise the efficiency of setting up fingerprint database and position.? Introduce a new particle swarm filter with adapt itself dynamically and extreme value variation,which means altering the inertia weight and accelerate factor dynamically with the different phase of the filter algorithm is,extreme value variation also.Works having been done all want to improve the performance of particle swarm filter and locate moving target accurately.Based on roughly location,acquire the best location of the target using N history position of target and the improved particle swarm filter algorithm.At last,we implement the location system for moving targets involving these improvements which have been presented before,and test the dominant modules' function.