Research On WIFI Indoor Positioning Teachnology Based On Location Fingerprint

With the development of mobile Internet and communication technology,location-based service has been applied widely in daily life,which makes positioning technology,especially indoor positioning technology,become more important.Wireless Fidelity(WIFI)indoor positioning technology based on localization fingerprint has become a research hotspot in the field of indoor positioning due to its intrinsic advantages,such as wide coverage,low cost and simple deployment.However,owing to the complex indoor environment and numerous interference,WIFI becomes strong time-variability and non-linearity in the process of indoor transmission,making us difficult to find its changing rules.Extreme Gradient Boosting(XGBoost)algorithm can efficiently fit nonlinear data,but it is rarely used in the indoor positioning.Therefore,this paper proposes an indoor location algorithm based on Candidate Construction Algorithm Based on Histogram-Extreme Gradient Boosting(CCH-XGBoost),and the main work is as follows:(1)The influencing factors in the WIFI transmission process are analyzed in this paper.Indoor environment is complex and changeable,and there are many factors that can influence WIFI signal.In this paper,the most important factors are analyzed to explain the reasons of their influence on WIFI transmission from the principle level.(2)An experimental platform is set up to collect WIFI signal strength data in the experimental area,and then study the properties of WIFI signal propagation process according to the data.Firstly,the properties of WIFI signal are studied in time and space dimensions,and it is concluded that WIFI intensity is time-varying,regular in a certain time,and irregular in space and inversely proportional to distance.Then the influence of terminal orientation it studied and finds that its influence on WIFI signal is negligible.(3)According to the properties of WIFI,this paper proposes a location fingerprint indoor location algorithm based on CCH-XGBoost.The CCH method is proposed from the properties of time and space dimension,which can construct Access Point(AP)Histogram in offline phase and quickly screen out the Candidate Reference Point(RP)of fingerprint data,improving the speed and accuracy of the location fingerprint positioning algorithm.Finally,XGBoost algorithm is used to train a binary classification tree model,which has a high fitting degree to the non-linear feature of WIFI and can effectively improve the accuracy of the localization algorithm.(4)The performance of the algorithm is verified and a nalyzed based on the public data set and the laboratory data set.The matching positioning results are verified on the public data set and compared with the nearest neighbor algorithm,XGBoost multi-classification algorithm and neural network algorithm.Then,the influence of two thresholds in the CCH on the positioning results is also analyzed.The Multi-RP fusion positioning results are validated on the laboratory data set and compared with the nearest neighbor algorithm.Numerous Experimental results show that the proposed algorithm is effective and has high positioning accuracy and low positioning error.The positioning algorithm proposed in this paper from the perspective of statistics and WIFI signal properties has been verified to be able to effective ly improve the positioning speed and accuracy of the location fingerprint indoor positioning,reaching the expectation of this paper.