JOINT LEAST-SQUARE SYNCHRONIZATION, CHANNEL ESTIMATION AND NOISE ESTIMATION

摘要

A method and apparatus for acquiring time synchronization, as well as a channel estimate and a noise estimate for a received signal are described. First, a burst of data which includes a training sequence is received. Then, groups of received training sequence measurements are created. For each of these groups, a set of linear equations is set up. Next, methods according to the present invention determine the least square estimate by solving the matrix equation H = (A?TA)-1*AT¿L for each of the groups. This least square estimate is performed in several steps which reduce the number of calculations to be performed. The term ATA is the same for each group and can be computed once prior to computing the least square estimates. The ATL terms are calculated separately for each group. In the process of calculating the ATL term for each group, the ATL term calculated for the previous group can be reused to reduce the total number of calculations. Once the A?TA and AT¿L terms are calculated, then the least square estimate for each group is determined. The least square estimate is then used to determine a noise estimate for each group. Again, mechanisms for reuse of certain calculations associated with the determination of the noise estimate that reduces the computational complexity associated therewith. Once the noise estimate is calculated for each group, the method determines which of the groups has the lowest valued noise estimate. That group, according to one exemplary embodiment, is chosen as the synchronization group. As an alternative, the group which has the greatest estimated signal-to-noise ratio can be selected as the synchronization group in this step. The least square estimate of the synchronization group is then identified as the channel estimate.