Bitwise Constellation Partitioning for Faster ML Demodulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-antenna wireless communication systems, existing demodulation techniques are complex and require significant processing resources, leading to increased complexity and processing time, especially when handling multiple streams and high-bit-rate modulation schemes, which are sensitive to channel impairments like noise.
Innovation Solution
The implementation of a multi-stream maximum-likelihood (ML) demodulation method that selects candidate data symbols, groups them into bit groups, and determines local optimum candidate values without calculating distance values, reducing the computational burden by using likelihood values based on global and local optimum candidate values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing demodulation techniques are used for multi-stream high-bit-rate modulation, then accurate demodulation can be achieved, but processing complexity and processing time increase significantly
Solution Approach 1:
The patent segments the demodulation process by dividing candidate data symbols into multiple groups based on bit positions. Instead of evaluating all candidate symbols simultaneously, the method processes them in organized groups, reducing the computational burden while maintaining demodulation accuracy for multi-stream high-bit-rate modulation schemes.
2Measurement precision
If existing demodulation techniques are used for multi-stream high-bit-rate modulation, then accurate demodulation can be achieved, but processing time increases significantly
Solution Approach 1:
The patent performs preliminary grouping of candidate data symbols based on their bit positions before the actual demodulation processing. By organizing candidates into groups in advance, the method reduces the time required for the subsequent evaluation and comparison steps, enabling faster processing while maintaining accuracy.
3Measurement precision
If maximum-likelihood demodulation is implemented for multiple streams, then demodulation accuracy is improved, but computational overhead increases
Solution Approach 1:
The patent reduces computational overhead by segmenting the maximum-likelihood evaluation process. Instead of computing distance values for all candidate symbols across all streams simultaneously, the method groups candidates by bit position and processes them in organized sequences, reducing the total number of computations required while maintaining ML demodulation accuracy.
Data Source
AI summary
A plurality of received data symbols is received, and a first received data symbol is selected from the plurality of received data symbols. A plurality of global optimum candidate values of a first estimated transmitted data symbol corresponding to the first received data symbol is determined for different given candidate values of second estimated transmitted data symbols corresponding to second received data symbols. Likelihood values for bits corresponding to the second estimated transmitted data symbols are calculated using the plurality of global optimum candidate values. All possible values of the first estimated transmitted data symbol are grouped into two or more bit groups, and a plurality of local optimum candidate values are determined for different bit groups. Likelihood values for bits corresponding to the first estimated transmitted data symbol are calculated using the plurality of global optimum candidate values and the plurality of local optimum candidate values.