WiFi Based Mobile Terminal Localization And Its Application

With the development of the mobile Internet,mobile phones and other smart mobile ter-minals have been integrated into people's daily life.When people use mobile Internet services,they generate a large amount of data,among which location data has a high value in use.It can be used to mine the spatial characteristics of the population,calculate the regional crowd density,and provide help for public safety incident warning,traffic monitoring,and regional planning It can also be used to determine individual contact and provide data support for screening close contacts of infectious diseases.The basis of location data is terminal positioning.Location information can be obtained by GPS technology outdoors,but in high-rise buildings or indoor environments,it cannot be obtained by GPS.In this case,WiFi positioning can be regarded as an effective way to obtain location information.This paper will study the specific application of mobile terminal positioning and location big data based on WiFi.The main research contents are as follows:Firstly,a WiFi location method based on a signal attenuation model is designed.Through two-step positioning,the problem of lack of signal source position in WiFi positioning is solved.This calculation method is simple and easy to be used in large-scale data.Secondly,for scenes requiring high positioning accuracy,a fine-grained WiFi fingerprint map generation model is proposed.The model takes into account the position noise of the crowd-sourced data,applies the coarse and accurate positioning data as a whole,and through the specially designed pixel lifting module and masking matrix,the performance of the fine-grained WiFi fingerprint map construction task exceeds all the other existing models and achieves the same accuracy as GPS in the positioning experiment.Thirdly,supplementing raw data with WiFi localization.Based on the grid model,we an-alyzed the crowd density of each area and designed a spatial-temporal feature extraction model which can integrate external factors.It can predict the crowd density of each area of the city according to the historical crowd density,weather,and date.The prediction is more accurate than that of the common time series prediction models such as LSTM and GRU.The prediction results can provide decision support for relevant government departments in problems such as traffic congestion,crowd gathering,and urban planning.Finally,we conduct a pioneering study on the role of digital contact in the prevention and control of infectious diseases,constructs a digital contact model in the context of COVID-19,and uses a machine learning algorithm to predict the disease probability of contacts.The results can help the government to narrow the search scope of high-risk contacts in epidemic preven-tion and control.