| Future wireless communication systems will need sophisticated techniques to operate reliably under typical radio propagation and interference conditions. Emerging multi-media applications demand high data rates with low delay, and limited battery capacity makes it essential to communicate extremely efficiently. Optimum receivers for unknown, time-varying channels encountered by wireless communication systems must jointly estimate the channel properties and decode the transmitted data. This is typically too computationally complex to implement. We propose a unified framework based on factor graphs for designing iterative receivers which approximate joint channel estimation and decoding using iterative algorithms. Once the appropriate graphical model for the receiver has been constructed, deriving the details of the algorithm is quite straightforward. In order to handle the problems introduced by continuous variables, we suggest canonical distributions, a general approach to simplifying the algorithms that allows their key features to be retained. Several receiver design examples illustrate our approach for common channel models.; Given the possibility of approximating joint channel estimation and decoding using iterative receivers, we would like to estimate the potential performance of the techniques in various situations. For channels with block memory, which are relatively easy to analyze and model a variety of frequency-hopping spread spectrum systems, we obtain bounds based on the channel reliability function, or error exponent. Both theoretical and numerical results confirm that the bounds behave qualitatively very much like iterative receivers with low-density parity check codes. Our results provide insight into the behavior of high-performance systems that use iterative receivers, especially considering the effects of channel memory on performance. |
