Research On Wi-Fi Indoor Localization Based On Location Fingerprinting

As the basis of many new technologies,positioning technology plays an important role.Although the outdoor positioning technology based on GPS is quite mature however,it cannot be applied to indoor environment directly because of its own limitations.It's difficult to find a perfect solution for the existing indoor positioning technology due to its high cost,narrow application range and low positioning accuracy.The development of indoor positioning is encountering great challenges.As an effective and feasible method to employ the signal characteristic data of Wi-Fi to determine the location,indoor positioning based on WiFi is widely used and the overall cost is not high.However,the limitation of Wi-Fi positioning lies in its heavy reliance on the positioning algorithm,which is directly related to the positioning accuracy.In order to improve the accuracy of indoor positioning,this paper investigated the research achievements of Wi-Fi indoor positioning technology at home and abroad in recent years,introduced the positioning principle of location fingerprint method,and studied the indoor positioning algorithm based on Wi-Fi.Meanwhile,it put forward Euclidean Distances and Hit Counts Weighted K-Nearest Neighbors(EDHCWKNN),Advanced Weighting based on Channels and Bandwidth(AWCB)and Unreliability Filtration(URF),which obviously improved the positioning accuracy.Specifically,this paper makes the following contributions.First,aiming at the limitations of the current WKNN algorithm,this paper proposes the introduction of "hit times" as a factor to be considered in the positioning algorithm.The EDHC-WKNN algorithm is introduced to calculate more scientific weights for data correction based on the relationship between Euclidean distance and hit times.Through comparison experiments with KNN algorithm and WKNN algorithm,it is concluded that the average error of EDHC-WKNN algorithm is respectively 22% and 17% lower than that of KNN and WKNN algorithm,and verifies the positioning accuracy of EDHC-WKNN algorithm.Secondly,combining the radio characteristics and the factors affecting the indoor positioning accuracy of Wi-Fi,this paper proposes the possibility that radio interference will result in the fluctuation of signal strength,and thus affect the final positioning result,and verifies it through experiments.The AWCB algorithm is introduced to estimate the radio interference degree by analyzing the channel and bandwidth configuration of the access point in the wireless environment,and then correct the European distance of the fingerprint point by weighting.Through the comparison experiment with the original algorithm,it is concluded that the mean error of AWCB algorithm is reduced by 1.70%,and the positioning accuracy is improved.Thirdly,based on the relevant MAC address specifications and the influencing factors of Wi-Fi indoor positioning accuracy,from the perspective of technical methods,this paper proposes an URF algorithm for credibility identification and filtering of access points in three aspects: visitor networks,access points that temporarily created and access points that created by unreliable devices.Experimental results show that the mean error of URF algorithm is reduced by 1.19%,that means,URF algorithm can identify and filter untrusted access points to a certain extent,reduce the positioning error,and further improve the positioning accuracy.