Maximumlikelihood detection for the multiuser codedivision multipleaccess (CDMA) channel is prohibitively complex. This paper considers new iterative multiuser receivers based on the expectationmaximization (EM) algorithm and related more powerful "spacealternating" algorithms. The latter algorithms include the SAGE algorithm and a new "missing parameter" spacealternating algorithm that alternately updates individual parameter components or treats them as probabilistic missing data. Application of these EMbased algorithms to the problem of discrete parameter estimation (i.e. data detection) in the Gaussian multipleaccess channel leads to a variety of convergent receiver structures that incorporate softdecision feedback for interference cancellation and/or sequential updating of iterative bit estimates. Convergence and performance analyzes are based on wellknown properties of the EM algorithm and on numerical simulation.
|Original language||English (US)|
|Number of pages||1|
|Journal||IEEE Transactions on Communications|
|State||Published - 1996|
All Science Journal Classification (ASJC) codes
- Electrical and Electronic Engineering