ADAPTIVE FORMATION METHOD OF TRANSITION MATRIX IN MARCOVIAN MOBILE COMMUNICATION ENVIRONMENTS

摘要

PURPOSE: A method for adaptive-tracking the transition matrix in a markovian mobile communication is provided to improve the quality of communication by modeling a channel coefficient provided according to the space into a markovian mobile communication channel, tracking its transition matrix, and selecting the channel model provided in the standard scheme according to the space non-related to the mobile station. CONSTITUTION: An approximation formula of Gaussian sum is obtained with respect to the initial distribution of a random state variable. A transition matrix is estimated out of the weight value of the Gaussian sum consisting of the observation value from the approximation formula and the error of the estimated value. The weight value of each Gaussian probability density function is detected in the Gaussian sum structure and if the approximation value is 1, only one Gaussian probability density function whose approximation value is 1 is selected out of several Gaussian functions in order to operate a Kalman filter. If there is no case where the weight value is 1, the state estimation amount is obtained by using several Kalman filters. The state estimation value and threshold value is compared to be identified.