Dynamic Space-Frequency-Division Multiple-Access over Frequency-Selective Slow-Fading Channels

This paper proposes a multiple-access (MA) framework: Dynamic Space-Frequency-Division Multiple-Access (DSFDMA), which is characterized by collectively and simultaneously considering the bandwidth assignment, subcarrier-clustering, power-allocation, beamforming vectors for both basestation and mobile terminals. This system design is optimized through mutual information maximization for broadband wireless systems over frequency-selective slow-fading channels. The DSFDMA conducts the transmission and reception processing in a cognitive manner, jointly dealing with the system requirements in terms of bandwidth efficiency, user-priority, and complexity-minimization for cases with diverse configurations of user-terminals. It is modeled and derived in a multiple-user MIMO-setup. And, the genetic algorithm (GA) optimization principle is used in the context of multiple-antenna multiple-user communications with multi-carrier signalings. The proposed scheme adapts to the channel states and interference status to maximize the overall system capacity as well as to maximize the individual user transmission capacity according to different priority settings. It also possesses valuable scalability and generality applicable to the heterogeneous terminals in both cellular and ubiquitous environment.