Indoor Positioning Based On Low-Cost Technology

Indoor location information has become a crucial data for many applications,such as travel navigation systems and advertising services,providing great convenience to people’s daily life,with the proliferation of mobile networks and smart devices.Since GPS(Global Positioning System)is ineffective in providing indoor location information,indoor positioning technology has gradually become a research hotspot.Compared with high-cost indoor positioning technology,low-cost indoor positioning technology has the advantages of being cost-efficient and widely applicable,so this paper focuses on low-cost indoor positioning technology.After detailed research and analysis of the advantages and disadvantages of various low-cost indoor positioning technologies,this study adopts an indoor positioning technology based on Wi-Fi fingerprint data as the research direction.Due to the problems of small dataset size,small indoor coverage area,and private fingerprint database in existing indoor positioning work based on Wi-Fi fingerprint data,the research results are not suitable for large indoor areas and difficult to compare algorithm performance.In addition,existing work mainly focuses on the specific coordinates of the current location for positioning,without comprehensive indoor positioning.Therefore,the research purpose of this paper is to comprehensively locate the building,floor,and coordinates of users based on the wide-coverage and public Wi-Fi fingerprint database.To address the shortcomings of low positioning accuracy in existing indoor positioning technologies based on Wi Fi fingerprint data and the problem of low precision and AP(Access Point)and RSSI(Received Signal Strength Indication)characteristics not being considered in the calculation process of the single KNN algorithm,this paper proposes two indoor positioning methods: indoor positioning method based on RSSI filtering and KNN algorithm,and indoor positioning method based on BPSO(Binary Particle Swarm Optimization),neural network,and KNN algorithm.The former proposes a filtering method based on the maximum value and interval of RSSI,uses multi-device standardization to solve the problem of non-uniformity of multiple-device RSSI measurement standards,and improves the positioning accuracy of the KNN distance measurement function.Based on BPSO,neural network,and KNN algorithm,the latter uses BPSO and Pearson correlation coefficient for feature selection to obtain the optimal AP combination,uses neural network to improve the accuracy of building and floor positioning,and uses KNN algorithm for coordinate positioning.The proposed indoor positioning method was experimentally evaluated under three criteria,and the best effect of the proposed indoor positioning method was compared with the average building positioning accuracy of the original algorithm,which increased by 1.71%,the average floor positioning accuracy increased by 4.32%,and the average coordinate positioning error decreased by 4.97 meters.