Systems with Iterative Channel Estimation

The estimation of channel parameters, such as the fading amplitude and the power spectral density of the interference plus noise, is essential to the effective use of soft-decision decoding. Channel estimation may be implemented by the transmission of pilot signals that are processed by the receiver, but pilot signals entail overhead costs, such as the loss of data throughput. Deriving maximum-likelihood channel estimates directly from the received data symbols is often prohibitively difficult. There is an effective alternative when turbo or low-density parity-check codes are used. The expectation-maximization algorithm provides an iterative approximate solution to the maximum-likelihood equations and is inherently compatible with iterative demodulation and decoding. Two examples of advanced spread-spectrum systems that apply the expectation-maximization algorithm for channel estimation are described and analyzed in this chapter. These systems provide good illustrations of the calculations required in the design of advanced systems.