Indoor Positioning Method And Implementation Based On Wi-Fi Fingerprinting

In recent years,indoor positioning has attracted much attention as one of the basic services of Internet of Things.Because the satellite signal is weak and cannot penetrate the buildings when it comes to the ground,we cannot use satellites to solve the location problem in indoor environment.Since Wi-Fi hotspots have been widely covered the indoor environment,users can access the Internet by carrying wireless communication devices from anywhere at any time.The indoor positioning method based on Wi-Fi fingerprinting is widely used because of its high positioning precision,no need to add extra hardware devices,simple and easy to do.This paper studies APs' influence on the indoor positioning method based on Wi-Fi fingerprinting,and analyzes how the probability distribution of the receiver signal strength is influenced by the line of sight condition and the quality of APs.Considering that the Wi-Fi signal is unstable and uncertain,and the contribution of different APs to the positioning are different,a fuzzy indoor positioning algorithm based on dynamic weights of APs is proposed in this paper.The uni-dimensional feature weighting based on kernel function is adopted to online assign weights for APs in different regions.The Multidimensional Scaling Growing Clustering(MSGC)algorithm is employed to online generate fuzzy rules.In order to maintain a compact and interpretable fuzzy rule base,the membership functions are merged according to the similarity of membership degree,and the consequent parameters are tuned by using the recursive linear-least-squares(RLS)algorithm.The proposed method is capable of not only weakening the influence of unstable APs on the final results,but also reducing the update frequency of the fingerprint data.Based on the indoor positioning algorithm proposed in this paper,an indoor location system is designed and implemented.In order to verify the proposed algorithm,the Wi-Fi fingerprinting is collected by the developed indoor positioning system under the actual environment.Comparison with the classical fingerprint location algorithm,the results demonstrate that the proposed method has advantages in positioning accuracy and anti-interference ability.