MIMO De-mapping via Successive Interference Cancellation and QR Decomposition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of de-mapping symbols in Multi-Input Multi-Output (MIMO) communication systems, particularly with increasing numbers of antennas, leads to high calculation complexity and performance limitations in existing detection methods like ZF, MMSE, and V-BLAST algorithms.
Innovation Solution
A method and apparatus that calculate multiple detected value sets and select the best one for de-mapping symbols, using a transmitted data symbol estimation unit and a detected value set decision unit, which reduces computational complexity and enhances estimation performance by employing Successive Interference Cancellation (SIC) with constellation point selection and QR decomposition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Maximum Likelihood algorithm is used for de-mapping symbols, then estimation accuracy is improved, but calculation complexity becomes excessively complicated
Solution Approach 1:
The patent divides the de-mapping process into multiple stages using Successive Interference Cancellation. Instead of processing all symbols simultaneously as in ML algorithm, it sequentially estimates and cancels interference from each transmitting antenna, breaking down the complex joint detection problem into simpler sequential steps while maintaining good estimation accuracy.
Solution Approach 2:
The patent calculates multiple detected value sets (more than the single optimal solution) and selects the best one based on correlation with received signals. This partial exhaustive search approach provides better estimation accuracy than conventional algorithms while avoiding the full computational burden of ML algorithm that would require searching all possible symbol combinations.
2Reliability
If V-BLAST algorithm is used for de-mapping symbols, then performance is improved, but calculation complexity increases due to iterative detection and inverse matrix calculation
Solution Approach 1:
The patent performs QR decomposition of the channel matrix beforehand to obtain orthogonal and upper triangular matrices. This preliminary transformation simplifies the subsequent detection process, eliminating the need for complex iterative inverse matrix calculations required by V-BLAST while maintaining good detection performance through the structured decomposition.
Solution Approach 2:
Instead of using iterative detection with nulling vectors as in V-BLAST, the patent inverts the approach by using QR decomposition to transform the channel matrix into a form where symbols can be estimated sequentially through simple matrix multiplication and subtraction operations, reducing computational complexity while maintaining reliability.
3Device complexity
If ZF or MMSE algorithm is used for de-mapping symbols, then calculation is simplified through linear detection, but estimation accuracy is limited compared to ML algorithm
Solution Approach 1:
The patent incorporates feedback by calculating multiple detected value sets and selecting the best one based on correlation with received signals. This feedback mechanism allows the system to refine its estimation by comparing predicted received signals with actual received signals, achieving better accuracy than conventional linear detection while keeping calculations manageable through the structured approach.
Data Source
AI summary
A method and an apparatus for de-mapping symbols in a Multi-Input Multi-Output (MIMO) system are provided. In the present invention, an estimated channel effect and a plurality of received data symbols are used to estimate a plurality of transmitted data symbols. A plurality of constellation points around an estimated value are selected as the detected values of the transmitted data symbol after the estimated value of one of the transmitted data symbols is calculated. Then, the detected values are used to calculate the detected values of remaining transmitted data symbols separately so as to obtain a plurality of detected value sets. Finally, a best one among the detected value sets is chosen as a result of de-mapping the received symbols. Therefore, only a few detected value sets are calculated and a better detected result is obtained, which possesses low complexity and high performance.