Research On Indoor Positioning Algorithm Based On WiFi Location Fingerprint

In recent years,with the development of multi-information fusion and the Internet of Things technology,smart phones and other wireless devices have become widespread on a large scale.Therefore,various services including indoor positioning have emerged.Outdoor positioning mainly relies on the GPS system,but GPS signals can be severely affected indoors and dense building group.So there emerged a large number of indoor positioning technologies.Many indoor areas have coved with Wi Fi due to the development of WLAN technology,so no additional beacons are needed for positioning.In the meantime location fingerprint positioning technology does not need to model complex indoor environments.Therefore,indoor positioning technology based on Wi Fi fingerprints has received widespread attention and study.However,due to the complex indoor environment and the susceptibility of Wi Fi signals,indoor positioning based on Wi Fi fingerprints is facing great challenges.The research in this paper is divided into the identification of buildings and floors and the positioning of same floor.Firstly,the recognition of buildings and floors based on deep learning is proposed.The Stacked Contractive Auto-Encoder network(SCAE)is added to the Deep Neural Network(DNN).In the offline phase,SCAE is used to extract the characteristics of the Wi Fi signal strength as the input to the DNN.In the online phase,the network is trained to identify buildings and floors.The SCAE-DNN network is robust to the fluctuations of input signal,so the recognition accuracy is high.For the positioning of same floor,an algorithm based on AP sequence to reduce the positioning area(Rdc-SRL)is proposed.By analyzing the sequence of AP sequences at different reference points,it is concluded that the relationship between the strength of Wi Fi signals between APs located at the reference point in the same area are relatively stable.According to this conclusion,the selected AP sequence is used to divide the positioning area into some parts.Firstly,located the part of the test point while positioning.The final position is determined by the Soft Range Limited k-Nearest neighbors(SRL-k NN)in the located part.By reducing the positioning area,it reduces the size of fingerprint database that the algorithm needs to traverse,the positioning accuracy is improved,and the computational complexity is reduced.At the same time,the problem that the signal spatial positioning algorithm such as k NN is not related to the actual space is effectively solved.In addition,the use of stable AP sequence positioning solved the problem of signal strength fluctuations.Buildings and floors identification performed using the public data set UJIIndoor Loc.The results show that the accuracy of buildings and floors using SCAE-DNN achieved 99.7%.The positioning in same floor is performed in an indoor positioning system that is actually constructed.Using the Rdc-SRL positioning algorithm,the algorithm achieves 75% of the error within 1m in the established experimental environment.