Turbo-Learning: A Novel Application to MIMO Wireless Communications

This paper briefly reviews the turbo-learning principle, thereby setting the stage for the description of a novel multiple-input, multiple-cutput (MIMO) wireless communication system. This new system, called Turbo-BLAST, builds on two key principles: The Turbo-learning principle The Bell-Labs Space-Time Layered (BLAST) architecture Experimental results are presented, demonstrating the high-spectral efficiency of Turbo-BLAST. A distinctive feature of the Turbo-BLAST receiver is its rapidly converging iterative nature. Indeed, it may be described as a nonlinear stochastic sampled-data feedback system. The paper identifies the development ofa mathematical framework for convergence analysis of the receiver as an open research problem.