Research On Wi-Fi Indoor Location Based On Dynamic APs

The subject comes from the project of Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences and the project is about ‘the development of mobile exhibition platform system of ‘Wisdom EXPO’, which used Wi-Fi based indoor location. Currently, the accuracy of Wi-Fi based indoor location is not satisfied which is the biggest obstacle for widely use. Aiming at this problem, we put forward the dynamic APs Wi-Fi based indoor location method. The theory and experiment show that the proposed algorithm can achieve better performance for Wi-Fi based indoor location.Firstly, this thesis reviews the present situation and the research methods at home and abroad. The basic concepts of the indoor positioning, trilateration based on Wi-Fi and the fingerprint positioning based on Wi-Fi were introduced as well as the study of the related content at home and abroad. Trilateration is a classic positioning algorithm with small amount of calculation but its deviation is high, and the fingerprint positioning algorithm has a high precision but a large amount of calculation. Studies have shown that combining the two algorithms will be a kind of new algorithm with better effect.Secondly, we establish the model of RSS. According to result of the field test, the distance path loss model of RSS was fixed. At the same time, we got the path loss attenuation n and the noise variance.Then we discuss the dynamic APs Wi-Fi based indoor location method in detail. On the basis of the analysis of traditional fingerprinting location method and trilateration and according to the experimental scene, we establish fingerprint database with the traditional method and the theoretical model of fingerprint positioning. Using the theoretical model above, we discussed the influence of neighborhood of the initial position, different matching algorithms and the numbers of APs(dimension of fingerprint vector) on the positioning accuracy and the amount of calculation, thus obtained the optimal positioning algorithm: firstly, according to the RSS from an AP, we estimated the system noise intensity, then we used trilateration to estimate the initial position and and obtained the maximum numbers of the matching reference points, and matched the fingerprint database with the 3NN matching algorithm and the optimal numbers of APs. At last, we compared the positioning accuracy and time of dynamic APs Wi-Fi indoor positioning method and the traditional fingerprint positioning method in different RSS noise cases.Finally, field test shows the proposed algorithm can obtain he positioning precision of 4.45 meters,within which the cumulative probability can reach 75%,when the standard deviation of RSS noise is lower than 7d B.This result can meet the requirements of the project.All in conclusion, we established the RSS model, fingerprint database and the theoretical model of fingerprint positioning. We propose the dynamic APs Wi-Fi based indoor location method. Engineering tests show this algorithm can obviously improve the positioning accuracy when the standard deviation of RSS noise is lower than 7d B. Predictably, with the development of the related hardware, positioning accuracy will be improved.