| 主权项 |
1. A method for iterative channel response estimation in a receiver of a multicarrier communication system that relies on forward-error-correction decoding, the method comprising:
selecting filter lengths for time and frequency domain channel estimation based on estimated time selectivity and frequency selectivity of said channel response, wherein said time selectivity and frequency selectivity are estimated using at least one of known pilot symbols and unknown data symbols; estimating, by an estimator, said channel response using said selected filter lengths by using available pilot symbols and a first set of data symbols, the first set of data symbols being selected as a fraction of data symbols that are more reliable than a remaining fraction of data symbols in a second set of data symbols; with respect to symbol positions corresponding to said second set of data symbols, estimating the channel response by interpolation based on estimated values of channel response obtained from said pilot symbols and said first set of data symbols; performing, by a decoder, forward-error-correction decoding using said channel response estimate to obtain improved estimates of coded bits; obtaining improved estimates of data symbols using said improved estimates of coded bits; calculating, by said estimator, an updated channel response estimate using a shorter filter length for at least one filter and a larger fraction of more reliable data symbols; repeating the steps of performing forward-error-correction decoding and obtaining estimates of data symbols based on said updated channel response estimate, and calculating further updated channel response estimates based on improved estimates of data symbols; and estimating noise power using the following steps:
estimating noise samples on positions of said known pilot symbols by subtracting from received noisy symbol samples a product of the corresponding estimated channel response and the pilot symbol values;estimating noise samples on positions of said first set of data symbols by subtracting from the received noisy symbol samples a product of the corresponding estimated channel response and estimated values of the first set of data symbols;estimating a power of said estimated noise samples associated with the positions of said first set of data symbols by either calculating a squared magnitude of said estimated noise samples associated with the positions of said first set of data symbols, or by squaring real and imaginary parts of said estimated noise samples associated with the positions of said first set of data symbols; andfiltering said estimated power over time with a first filter length and over frequency with a second filter length. |
