Maximum likelihood channel estimation with noise of unknown power spectrum

In this paper, we present an algorithm for obtaining maximum-likelihood (ML) channel estimates when the baseband noise is colored with an unknown power spectrum. The algorithm jointly estimates the channel and the spectrum of the noise in an iterative manner. We use the EDGE (enhanced data rates for global evolution) air interface to evaluate the impact of the proposed channel estimation technique in both interference dominated and thermal noise dominated situations. Simulation results show that the algorithm is very effective in suppressing the adjacent channel interference. For example, about 5 dB reduction in C/I required to achieve 10% block error rate at coding rate of 0.76 can be achieved