Two-step multi-user indoor location estimation based on Bayesian approach

This paper proposed a two-step indoor location estimation algorithm. A Bayesian approach is adopted to incorporate prior PDF of the concerned variables to obviate the use of training sequences. Then the preliminary estimated parameters were combined with the path loss model. The estimated location was modified in its neighborhood to achieve greater accuracy. Considering the great difficulty in acquiring training sequences in actual conditions, this algorithm is particularly useful. Besides, if the number of users increases, the previously estimated posterior PDF is utilized to estimate the locations of the newcomers alone, rather than computing all the users altogether again. This property greatly saves the amount of time consumed in location estimation