Efficient WiFi Localization Of Single AP Based On Environment Learing

With the development of Internet of Things technology,applications such as social networking,shopping navigation and smart home have become an indispensable part of people's lives,and accurate target positioning is a key element of these applications.The traditional target location method is mainly based on GPS technology and fingerprint technology.GPS-based positioning method can only achieve fast and accurate positioning in outdoor.In complex indoor multi-path environment,GPS accuracy is greatly reduced or even impossible to use;fingerprint-based positioning method requires a large amount of manpower to periodically collect fingerprint database.And its positioning accuracy is easily affected by the environment.Therefore,the low-cost precision positioning technology has become an urgent problem to be solved.In recent years,due to the low cost of Wi-Fi devices,they have been widely deployed in most indoor public places and homes.Therefore,indoor positioning based on Wi-Fi devices has attracted extensive attention in the academic community.At present,most indoor positioning technologies based on Wi-Fi wireless signals either sacrifice positioning accuracy at a low cost or sacrifice the cost for positioning accuracy.Therefore,in order to achieve low-cost high-precision indoor positioning,an indoor positioning method based on neural network band splicing is proposed.However,in the process,we are faced with the following challenges: 1)The CSI of the original data is large in error due to the offset load caused by the hardware crystal oscillator and the clock offset of the transceiver end,and cannot be directly used;2)Frequency hopping brings additional frequency hopping drivers on the client side,and the client's devices usually use smart phones,smart watches and portable devices equipped with Wi-Fi chips.Additional drivers will increase the energy consumption of the device;3)Data communication is an important part of all applications,and constantly switching channels will seriously affect the stability of the ongoing data communication of the clien.How to use only a single AP without excessive human intervention is a difficult problem to be solved.In order to solve the above problems,the main specific research contents of this paper are as follows:(1)Analyze the offset cause and offset effect caused by the hardware clock offset to the CSI measurement data,and obtain the real and effective CSI data under the channel by preprocessing the data.(2)method for learning channel state information based on a neural network is proposed.By extracting channel state information CSI features within a plurality of channels under coherence time,frequency band splicing is realized,and a virtual high bandwidth signal is constructed.The dynamic variability of the indoor environment and the influence of multipath are the decisive factors affecting the indoor positioning system.If the response time of the system model is too long,the state information characteristic curve of the channel changes,which will invalidate the model.In response to this challenge,the efficiency of the neural network can be well obtained based on the characteristics of the channel state information in the channel stabilization time and updated in time;the splicing phase and the positioning phase are separated,and the client still uses single channel communication.This will not bring extra hopping load to the client.(3)A multi-factor positioning system based on frequency band splicing is proposed.The multipath and non-line-of-sight paths in the indoor environment seriously affect the accuracy of positioning.When the non-line-of-sight path and the line-of-sight path are similar,the low-bandwidth resolution The rate will not be recognized.Parallel clients can achieve higher virtual bandwidth by using single channel communication,which will not affect the data communication being transmitted,and can obtain higher resolution in both time and frequency domains.This paper uses a comprehensive The multi-factor clustering method eliminates the outliers of the line-of-sight path and achieves high-accuracy and highprecision positioning results.A large number of experiments have proved the effectiveness of the indoor positioning method proposed in this paper.In the complex indoor environment,this paper can achieve an average median error of 0.7 m for the target.