Reduced Complexity GA/RBF Multiuser Detectors for DS/CDMA Communications

L.M. San José Revuelta (Spain)


CDMA, Low complexity multiuser detection, MicroGenetic Algorithm, RBF networks


This paper presents two novel structures based on modified Natural Computation techniques for multiuser detection in stationary and fading channels. Both methods share the common property of low complexity. First, a -GA (a mod ified low complexity Genetic Algorithm with weak com putational load), whose parameters are on-line adjusted by the entropy of the population fitness, is presented. Sec ond, a Mahalanobis based Radial Basis Function receiver is analyzed. Both algorithms are shown to have moderate computational load as the number of users increases, and offer an interesting alternative to previous suboptimal al gorithms whose performance is frequently subject to the near-far problem and multiple access interference degrada tions. Their performance and characteristics are compared with those of classical GAs- and NNs-based detectors, and with traditional multiuser detectors, such as the matched filter, the decorrelator and the MMSE detectors.

Important Links:

Go Back