发明名称 |
SYSTEM AND METHOD FOR TREE-SEARCH ENHANCEMENT BY METRIC PREDICTION BASED ON INCOMPLETE PATHS IN SOFT OUTPUT MIMO DECODER |
摘要 |
A method and system for soft output multiple-input-multiple-output (MIMO) decoding may include generating a tree-graph based on: MIMO rank, number of bits per layer, and type of modulation, wherein the tree-graph comprises a root node, leaf nodes, nodes, and branches connecting the nodes; performing sphere decoding by determining a radius covering a subset of nodes within said tree-graph; managing, based on the sphere decoding, tables comprising metrics and counter metrics usable for log likelihood ratio (LLR) generation; predicting, based on a specified prediction scheme, counter metrics for paths in the tree-graph that comprise nodes and branches out of the determined radius; and updating the tables comprising the counter metrics with the predicted counter metric, in a case that the predicted counter metrics are better in maximum likelihood terms than the determined counter metrics. |
申请公布号 |
US2015222457(A1) |
申请公布日期 |
2015.08.06 |
申请号 |
US201414341869 |
申请日期 |
2014.07.28 |
申请人 |
Ceva D.S.P. Ltd. |
发明人 |
KAPLAN Zeev;Dvoretzki Noam;Hai Eitan |
分类号 |
H04L25/03;H04B7/02 |
主分类号 |
H04L25/03 |
代理机构 |
|
代理人 |
|
主权项 |
1. A method of soft output multiple-input-multiple-output (MIMO) decoding, the method comprising:
generating a tree-graph of bits based on: MIMO rank of a receiver, number of bits per layer in the graph, and type of modulation, wherein the tree-graph comprises a root node, leaf nodes, nodes, and branches connecting the nodes; performing sphere decoding by determining a radius covering a subset of nodes within said tree-graph; managing, based on the sphere decoding, tables comprising metrics and counter metrics usable for log likelihood ratio (LLR) generation; predicting, based on a specified prediction scheme, counter metrics for paths in the tree-graph that comprise nodes and branches out of the determined radius; and updating the tables comprising the metrics and the counter metrics with the predicted counter metric, in a case that the predicted counter metrics are better in maximum likelihood terms than the determined counter metrics. |
地址 |
Herzlia Pituach IL |