Research On Indoor Positioning Technology And Application Based On WLAN

With the rapid development of mobile internet technology and smart phone,The social networks,shared trips,traffic navigation,emergency rescue and other applications which are based on location service has shown great market prospects,and also in indoor environments such as supermarkets,hospitals,guilds,prisons,theatres and libraries.A large number of indoor location services have sprung up too.Indoor location service is the foundation of truly realizing the internet of things,and It will become the next trillions of blue sea market,with great social,economic and scientific and technological strategic value.How to achieve indoor location service with high accuracy and low cost has become a research hot spots both at home and abroad.The indoor positioning technology based on WLAN has become the key point in the research of indoor positioning technology because of the wide deployment of Wi-Fi network,low cost and universal support of the smart phone.However,the indoor structure is complex and changeable,the terminal acquisition RSS is different,the movement state of the smart phone while positioning is different and the unified indoor location application model is lack.All these problems greatly restrict the development of indoor location technology based on WLAN.On the basis of the comprehensive research and analysis of the development of indoor positioning technology,this dissertation has carried out extensive and in-depth research on some key problems existing in indoor positioning,and obtained the following research results:(1)Based on signal propagation model,ranging and positioning technology has larger error,and it is difficult to collect and maintain fingerprint data by using the fingerprint matching technology.In order to solve these problem,An indoor location method integrating indoor signal propagation model and fingerprint matching is proposed.Based on the spatial distance between AP and the reference points.The method constructs rss fingerprint samples by using indoor signal propagation model and the AP rss distribution characteristics at the reference point.and then rss fingerprint samples are transformed into gray image samples,and finally,the convolutional neural network(CNN)is used to construct the classification location model.This method transforms the traditional rss fingerprint matching into image matching,and greatly reduces the workload of fingerprint collection and the difficulty of maintenance,and the indoor positioning accuracy of an average of 2.18 meters is obtained.(2)A mobile location algorithm based on convolution neural network and smart cell phone Sensor is proposed to solve the problem of location error introduced in indoor mobile location scene due to the difference of Wi-Fi chip and the change of positioning terminal movement state.Through experiments,it is found that different types of smartphones have received the different AP rss while walking in the same route,but the change trend of continuous reference point is very similar.By using this feature,a number of continuous reference points rss on the movement route is transformed into image samples,and a CNN is used to train the classification location model.In the online positioning stage,the current motion state and direction of the terminal are judged by the acceleration Sensor and gyroscope Sensor of the smart phone,and the AP rss received by the positioning terminal is transformed into image data,and the classification model is used to show the position.Using the smart phone to test in the indoor environment,this method has achieved an average location error within 3 meters.It can support different types of terminals well,and the positioning accuracy is better than the fingerprint matching location algorithm based on the nearest neighbor.(3)For some smart phones that do not open the AP signal intensity scan function or the terminal user does not install the location application,a reverse indoor location method based on programmable AP is proposed.First,a highly efficient and stable rss perceptron(Sensor)is designed based on Samsung S3C2440A chip and Broadcom BCM4331 communication chip.Sensor uses DD-WRT firmware to effectively capture the communication data under 802.11 environments and extract the rss.Secondly,the deployment and optimization methods of Sensor in the positioning environment are designed,and based on the Voronoi diagram,the location area is decomposed into several regions.In each region,the spatial propagation model of the signal is continuously optimized through the optimization mark between Sensors.Finally,the three side location algorithm is perfected to calculate the position of the smart phone.The average error is 2.76 meters,which is better than the three edge location algorithm without using space signal optimization.This method can effectively support different types of intelligent terminals,and has the advantages of low hardware cost,flexible deployment and high positioning accuracy.(4)In view of the problem that the application business of indoor positioning is not clear and flexible,a cloud service application model for indoor positioning is proposed.The model defines the related roles,service framework and location algorithm scheduling method of indoor location,defines the correlation of the related roles,simplifies the development process,and is beneficial to the rapid deployment and popularization of the indoor positioning application.(5)In the laboratory environment,the main functions of the system,the location model and the algorithms proposed in this dissertation are tested.The experimental results show that the algorithms,the positioning application model and the prototype system can effectively fulfill the indoor location needs.At the same time,based on Neu-iLocation,a typical application scene of fire rescue based on indoor location service is built,and a fire rescue prototype system is designed and implemented.The prototype system verifies the effectiveness of the related research results,and the significance of the research again.