A gradually locating method of indoor locating estimation based on likelihood

Locating using wireless signal is a popular field now, and the real time indoor locating is a difficult problem for its complexity and sensitivity to environments. This paper proposes a gradually locating method based on Euclidean distance and the maximum likelihood, which maintains both Euclidean distance's robusticity and the maximum likelihood's high precision under complex environments. To reduce the number of supervised vertices in training data required by the grid-matching algorithm, this paper also presents an interpolation method based on the received signal strength (RSS) model in the local area, which successfully simulates the real signal distribution on the interpolation point. By using the above method, we can obtain unbiased locating result, and the locating can approach to the real position steadily as the amount of signals increases. Copyright 2007 ACM.EI