This project investigates feasibility of transmitter optimization for rapidly time varying fading channels encountered in wideband mobile radio communication systems. Adaptive transmitter optimization methods have the potential to significantly improve performance in wireless channels by adjusting the transmitted signal-to-fading conditions. However, rapid channel variation makes application of these closed-loop techniques difficult in wireless systems. Previous investigations demonstrate that it is not sufficient to utilize the outdated Channel State Information (CSI) in adaptive transmission algorithms when vehicle speeds are significant. Prediction of the channel coefficients several tens-to-hundreds of symbols ahead is essential to realize these methods in practice. A novel adaptive, long-range fading channel prediction algorithm (LRP) was previously developed by the PI and her collaborators for narrowband wireless channels. In this research, the long range fading prediction capability is extended to frequency selective fading channels with antenna arrays. These methods are utilized in adaptive modulation and coding techniques for multicarrier and multiple antenna systems and in novel precoding methods and joint transmitter/receiver optimization techniques for Multiple Input Multiple Output (MIMO) wideband mobile radio systems. This research contributes to the development and realization of adaptive transmission and precoding methods that are essential in reliable high rate wireless communication. Moreover, collaboration with industry benefits graduate students involved in the project. Broader participation of underrepresented groups is enhanced since the PI and the graduate students are females. The infrastructure for research and education is enriched by incorporating the methods investigated in this research in senior and class projects and by integrating with wireless communications activities at NC State. The first component of this project focuses on utilization of prediction methods in interference cancellation for the MIMO channels that arise in multicarrier, multiple antennae and the downlink of multiuser Direct Sequence Code Division Multiple Access (DS/CDMA) channels. While the sources of interference depend on the application, these systems can be described with a single underlying MIMO model. Novel Tomlison-Harashima precoding methods are developed for this model. These methods are expected to improve significantly on previously proposed linear precoding techniques. Moreover, they distribute computational resources between the transmitter and the receiver as dictated by the application. For the downlink of the DS/CDMA channel, novel low complexity linear techniques that have the potential to significantly simplify precoding at the base station and detection at the mobile station are investigated. The proposed interference cancellation techniques are combined with adaptive modulation and coding. Since these precoding and adaptive transmission methods depend on accurate CSI in the transmitter, reliable LRP and the underlying prediction error statistics are essential in the design and the performance analysis of the proposed methods. In the second component, joint prediction for multiple subcarriers in multicarrier systems and antenna array elements in multiple antenna channels is investigated. Novel linear and nonlinear approaches to long-range prediction are explored. A statistical model of the prediction error is developed and utilized in the design of reliable adaptive modulation and coding methods. Flexible prediction methods that trade spectral efficiency for complexity and feedback requirements are explored.
