Indoor Localization Method Based On Multiple Data Sources

Nowadays, with the quick development in data services and multimedia services, the demandfor navigation and localization is becoming more and more.Specifically in some complex indoorroom-level environments, such as the company, hospital, tea room, supermarket, market,museum, cinema and any other environments, people often need to abtain the locationinformation of the mobile terminal or their holders, facilities or the things in the room.Depending on the indoor and outdoor different positioning environments, nowadays wirelesslocation technology can be divided into indoor positioning technology and outdoor positioningtechnology.Indoor positioning technology name suggests, refers to a way to get the objects inindoor location technology. It has a broad application future in military and civil domains. Theknown localization technology main include wireless local area networks, light tracking sensernetwork positioning technology, infrared ray positioning technology, radio frequencyidentification positioning technology (RFID). Therefore, sensor technology and wireless localarea networks technology with their own advantages can be applied to indoor positioningmethod.The thesis combines sensor networks and wireless local area network (WLAN) to localizeand track the single target indoor. First, the thesis describes the background of using sensornetworks and WLAN technology for indoor location, and then discusses the received signalstrength data for the conduct of activities, using light intensity sensor and wireless LAN in theindoor environment. Specifically, we use the light intensity sensor to abtain the initial positioningof the person in the room, and then use the wireless LAN received signal strength data for theprecise positioning. Among many machine learning algorithms, the neural network hasadvantages such as being adept in resolving nonlinear mapping problems and precisionadjustable This thesis uses the method for persons to obtain location information based on thereceived signal strength and the BP neural network. We propose an indoor localization modelfrom three-layer BP neural network which has one input layer one hidden layer and one outputlayer. The BP neural network use signal intensity as its input vector and is a kind of parallelstructure.The output signal of hidden layer is transimited to output layer, througth the outputlayer the result is outputted, finally gives an output result, which is the coordinate pointsindicating the precise location of the indoor personnel. Finally, the newly proposed model is simulated in wireless LAN evrionment in terms of IEEE802.11b standard, and the results showthat the application of the model can effectively give locaion of the person in the room...