Linear precoding for multi-input systems based on channel estimate and channel statistics
AssigneeThe Board of Trustees of the Leland Stanford Junior University
InventorsMai H. Vu, Arogyaswami J. Paulraj
Life of this patent
Estimating the term…
Loading patent…
AssigneeThe Board of Trustees of the Leland Stanford Junior University
InventorsMai H. Vu, Arogyaswami J. Paulraj
Estimating the term…
A method, processor and system for linear precoding for a multiple-input communication channel described in terms of an effective mean and an effective correlation is provided. The effective mean and correlation can be the sample mean and correlation of the channel, or they can also include an estimate of the channel and a quality parameter of the estimate. The precode is derived by minimizing a Chernoff bound on the pairwise error probability. It is assumed the channel code is a matrix block code, which includes spatial multiplexing and space-time block codes (STBCs) as special cases. In some cases (e.g., an orthogonal STBC) the codeword separation matrix is a scaled identity matrix. In these cases, the precode that minimizes the Chernoff bound can be determined analytically. It may be necessary to perform a “dynamic water-filling” procedure in order to satisfy constraints on the optimization. In other cases, the codeword separation matrix is not a scaled identity matrix. In such cases, various precodes which approximately minimize the Chernoff bound are provided.
A method for transmitting information comprising:
Dependencies are inferred from the claim text.