Research On WiFi Location With High Accuracy Based On Machine Learning

As the Location Based Service plays more and more important roles in the life of people, Wireless Location is becoming a hot spot in the area of location technology. The research of this thesis design and complete one kind of wireless location technology based on location model with the Wi Fi technology which is widely used.The main work is as follows:1 This thesis has done the research in the ensemble learning and CCA. On the basis of it, this thesis proposes a new method named EMCCA, which combines MCCA and ensemble learning. Firstly, this thesis study and analysis the CCA and DCCA, GCCA, MCCA, which is based on CCA. In order to improve the accuracy and generalization of learning machine, we divide the sample datasets and do feature fusion on these datasets and combine with ensemble learn. In the experiment of UCI datasets, include Ionosphere, Survivor, Heart, and MFD, the results of experiments show that, comparing with DCCA and GCCA, the EMCCA has more accurate and better generalization performance.2 In order to get Visual performance of the data, which has three or more dimension, when getting classification with SVM, this thesis propose a visual method of selecting parameters of SVM. Based on the theory of Visualization and Support Vector Machine, this thesis propose an algorithm to get classification interface in any dimension, and then propose another algorithm to get visual image in any dimension classification interface. With the help of Visualization and Human-Computer-Interaction, the new method, this thesis proposed, can help get better parameters.3 This thesis proposes a new method and implements a Location system based on Wi Fi, which is reverse and synchronous. The word of ‘reverse' is meaned that during the location, it is necessary to turn on the access point for the sensing points, which have been set in the environment before. The word of ‘synchronously' is mean that the sensing points are controlled by the synchronous controller, and synchronously scan the access points, upload the result to the synchronous controller synchronously. The system analysis the data with the visual method and get the classification with EMCCA. The results of experiments show that the location system is more stable, lowly cost, easy to use, and safe with location information of users.