Research On Wi-Fi Indoor Positioning Technology Based On Location Fingerprint

With the increasing demand for location-based services (LBS), wireless positioning technology has become a hot research topic. In consideration of low cost, acceptable accuracy and instrument penetration, Wi-Fi positioning technology has gradually become a major wireless positioning technology. Wi-Fi positioning technology based on location fingerprint can use only an AP for positioning and has been widely used. Therefore, this thesis focuses on Wi-Fi positioning technology in different indoor environments.Due to the complexity of the indoor environment, three environment cases have been configured to represent different indoor environments. Case one is a section of a narrow corridor, which is used to represent the corridor environment. Case two is a classroom, which is used to represent the office environment. Case three is a corner, which is used to represent the corner environment.Firstly, based on K-nearest neighbor algorithm and VaS algorithm, two new matching algorithms were proposed. One uses Manhattan distance as distance metric and calculates the mean of the K-nearest reference locations as the final positioning result. The experiment results show that the positioning accuracy increases up to 10% compared to KNN algorithm. The other uses Mahalanobis distance as the distance metric and calculates the mean of the K-nearest reference locations as the final positioning result. The experiment results show that the positioning accuracy increases up to 15% compared to KNN algorithm.Secondly, we studied the relationship between the number of AP and positioning accuracy. Different experiment cases using the same number of AP show different positioning accuracy, since 43% of the test cases have positioning errors within 5 m in case one but 100% in case two and case three using an AP. Moreover, the number of AP shows different effect on positioning accuracy in different cases. The test cases with positioning errors within 5 m increased by 20% in case one when the number of AP increased by 3, but only increased by 5% in case two and case three.Finally, we studied the relationship between the location of AP and positioning accuracy. In experiment case three, we placed three APs in the form of triangle deployment and a line deployment. The experiment results show that the triangle deployment performs better than a line deployment, where the average positioning error decreased by 0.5 m and the minimum positioning error decreased by 1 m.Overall, Wi-Fi indoor positioning technology is affected by many factors. Therefore, the APs for positioning should be used as much as possible in the scattered positions. Manhattan distance can be used as an index of matching since it has good robustness. When there are no limits to complexity, the size of the fingerprint database should be timely increased.