Reduced-Complexity MLD for SC-FDMA Symbol Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current detection algorithms, such as Maximum Likelihood Detector (MLD), are inefficient for processing SC-FDMA data streams due to the exponential growth of calculations required, making it infeasible to apply full MLD complexity to SC-FDMA, especially in high-data-rate scenarios like 20 MHz LTE uplinks with multiple 64-QAM streams.
Innovation Solution
The method involves obtaining initial estimates of symbols using a non-MLD algorithm, reducing the search space complexity, and applying a reduced-complexity Maximum Likelihood Detection (RC-MLD) algorithm only to a limited search window within the digital communication input, leveraging the concentration of symbols in time to model and extract a subset of symbols efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full Maximum Likelihood Detection (MLD) is applied to extract symbols from SC-FDMA data streams, then symbol extraction accuracy is improved, but computational complexity grows exponentially making it infeasible
Solution Approach 1:
The patent segments the SC-FDMA signal processing into distinct stages: FFT transformation, time-domain equalization, and symbol extraction. By dividing the processing chain, the patent applies different detection complexities to different signal components, avoiding full MLD complexity while maintaining accuracy for critical symbols.
Solution Approach 2:
The patent applies reduced-complexity detection methods (such as linear equalization or simplified detectors) to a subset of symbols rather than all symbols. This partial application of complex detection is sufficient to achieve the desired performance while dramatically reducing computational burden compared to exhaustive MLD.
2Device complexity
If reduced-complexity detection algorithms are used instead of full MLD, then computational complexity is reduced, but symbol extraction accuracy deteriorates
Solution Approach 1:
The patent applies different detection qualities to different parts of the signal. Critical symbols that require high accuracy are processed with more sophisticated detection methods, while less critical symbols use simpler detection. This localized quality adjustment optimizes the balance between complexity and accuracy.
Solution Approach 2:
The patent performs preliminary signal processing steps (FFT transformation and time-domain equalization) that prepare the signal to make subsequent symbol extraction easier and more accurate. These preliminary actions transform the signal into a form where reduced-complexity detectors can achieve near-MLD performance.
3Productivity
If MLD search space is reduced by limiting time window, then processing speed is improved, but symbol detection accuracy may be compromised
Solution Approach 1:
The patent employs dynamic adjustments to the detection process, including adaptive equalization parameters and flexible search window sizing. The system can dynamically adjust the time window length and detection complexity based on signal conditions, maintaining accuracy while optimizing processing speed for different scenarios.
Data Source
AI summary
A method and system are described for applying MLD detection to SC-FDMA data streams. Embodiments obtain initial estimates of at least some symbols encoded within an SC-FDMA data input using a non-MLD algorithm, and then use the initial estimates to reduce the search space and the MLD complexity. In addition, the reduced complexity MLD (“RC-MLD”) is applied only to a limited search window within the digital communication input, thereby modeling only a subset of the symbols included in the digital communication input. The MLD complexity is thereby reduced sufficiently for an RC-MLD decoder to simultaneously model and extract a plurality of the symbols included in the data input. Symbols initially decoded by the non-MLD algorithm can be subtracted from the data input in a serial interference cancelling module. This cancellation can be applied successively in a turbo-loop.