Linear prediction based multiuser detection for multicarrier CDMA communications systems

Multicarrier (MC) CDMA systems have been developed and rapidly gained popularity as they capitalize on both MCM (multicarrier modulation)´s resilience to multi-path fading and direct-sequence (DS) CDMA´s simplicity in multi-access and synchronization design to mitigate both multiple access interference (MAI) and the inter-symbol interference (ISI) caused by time dispersive channels with the aid of cyclic prefix (CP). Nevertheless, the elimination of ISI in the MC-CDMA communications systems is at the cost of significant loss in bandwidth efficiency. Additionally, due to the time varying nature of the multipath channel, the training sequence needs to be transmitted periodically, causing further loss of channel throughput. Therefore, blind channel estimation methods which avoid the use of training sequence or even the CP are more attractive due to the increasing demand for high-bit-rate transmission. In this paper, a blind linear prediction based multiuser detection method which avoids the use of CP is proposed for MC-CDMA communications systems. Compared with some existing blind methods, our newly proposed method has better performance. Computer simulations are presented to verify its effectiveness.