Neural Network Peak Reduction Tone Generation for OFDM PAPR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional orthogonal frequency-division multiplexing (OFDM) waveforms in 5G NR suffer from high peak-to-average power ratio (PAPR), leading to power amplifier back-off and degraded efficiency due to power amplifier nonlinearity, with traditional signal processing methods lacking a meaningful relation between data tones and peak reduction tones (PRTs).
Innovation Solution
The use of machine learning-trained neural networks to map data tones to PRTs for transmission, where a PRT neural network and a receiver neural network pair are trained to specifically pair data tones and PRTs, enabling the receiver to reconstruct data tones accurately by leveraging information embedded in the PRTs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional OFDM waveform is used, then data transmission is achieved, but peak-to-average power ratio becomes large causing power amplifier back-off and efficiency degradation
Solution Approach 1:
The patent segments the OFDM waveform into data tones and peak reduction tones (PRTs). The PRTs are separate auxiliary tones that do not carry data but are specifically designed to reduce peak values. This segmentation allows the system to maintain data transmission while separately addressing the PAPR issue through dedicated PRTs that cancel peaks without interfering with data tones.
Solution Approach 2:
The patent introduces peak reduction tones (PRTs) as intermediary elements that mediate between the data tones and the power amplifier. These PRTs act as a buffer that absorbs the peak excursions caused by data tones, preventing the power amplifier from entering nonlinear operation. The PRTs are calculated based on the data tones and are added to the waveform before power amplification.
2Object-generated harmful factors
If traditional signal processing methods are used to generate PRTs, then peak reduction is achieved, but meaningful relation between data tones and PRTs is lacking reducing demodulation accuracy
Solution Approach 1:
The patent implements a feedback mechanism where the receiver detects the PRTs in the received waveform and uses this information to reconstruct the original data tones. The PRTs are not simply removed but are actively used as feedback signals that contain information about the peak reduction process. This feedback loop ensures that the relationship between data tones and PRTs is preserved and exploited for accurate demodulation.
Solution Approach 2:
The patent makes the PRTs serve multiple functions: they reduce peaks to improve power amplifier efficiency, and simultaneously carry information that aids in accurate demodulation. By designing PRTs with dual functionality, the system eliminates the trade-off between peak reduction and demodulation accuracy. The PRTs are generated using neural networks that optimize both peak reduction performance and demodulation accuracy simultaneously.
Data Source
AI summary
Various embodiments include methods for demodulating wireless transmission waveforms to reconstruct data tones, performed in receiver circuitry of a wireless communication device. The methods may include receiving time domain wireless transmission waveforms that include time domain peak reduction tones (PRTs) that were generated by a set of PRT neural networks using time domain data signals derived from frequency domain data tones in another wireless communication device, reconstructing the time domain data signals from the time domain wireless transmission waveforms using a receiver neural network that has been trained in conjunction with the set of PRT neural networks to generate a time domain reconstruction of data signals, and transforming the time domain reconstruction of data signals to a frequency domain reconstruction of data tones.