MIMO-OFDM PAPR Reduction via Iterative ANN Precoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Massive MIMO-OFDM systems face significant challenges due to high Peak to Average Power Ratio (PAPR) caused by large envelope fluctuations, leading to severe nonlinear distortions when signals pass through high power amplifiers.
Innovation Solution
The proposed solution involves an iterative PAPR reduction scheme based on Artificial Neural Network (ANN) technology, which generates sequences of OFDM symbols, determines a processing matrix, adjusts and clips reference values of parameters in a signal generation model to minimize PAPR while preserving signal information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If OFDM signals are transmitted through high power amplifiers, then communication capacity and speed are improved, but severe nonlinear distortions occur due to large envelope fluctuations
Solution Approach 1:
The patent applies preliminary action by performing PAPR reduction processing on the OFDM signal before it enters the high power amplifier. The signal is processed through a signal generation model with adjustable parameters that are optimized in advance to minimize envelope fluctuations, preventing the generation of harmful nonlinear distortions while maintaining the signal's ability to achieve high communication capacity
Solution Approach 2:
The patent implements parameter changes by adjusting the parameters of the signal generation model (such as precoding matrices and signal weights) to transform the signal characteristics. By iteratively optimizing these parameters, the signal's envelope fluctuations are reduced, allowing the high power amplifier to operate in a more linear region while preserving communication performance
2Object-generated harmful factors
If PAPR reduction processing is applied to OFDM signals, then envelope fluctuations are reduced, but signal information may be lost
Solution Approach 1:
The patent employs feedback by iteratively adjusting the parameters of the signal generation model based on the PAPR of the generated signal. The process continuously monitors the signal characteristics and refines the parameters to achieve optimal PAPR reduction while preserving signal information, balancing the reduction of envelope fluctuations with information preservation
Solution Approach 2:
The patent applies dynamics by using an adaptive signal generation model where parameters can be dynamically adjusted during the signal generation process. This allows the system to flexibly optimize the signal characteristics in real-time, reducing envelope fluctuations while maintaining signal integrity through iterative parameter optimization
3Object-generated harmful factors
If iterative parameter adjustment is performed to reduce PAPR, then PAPR is reduced to very low levels, but processing complexity increases
Solution Approach 1:
The patent manages processing complexity by systematically adjusting parameters of the signal generation model in an iterative manner. By focusing optimization on key parameters (such as precoding matrices and signal weights) rather than all possible signal characteristics, the system achieves significant PAPR reduction while keeping the processing complexity at manageable levels
Solution Approach 2:
The patent uses a signal generation model that generates signals with desired characteristics through parameter optimization. By working with the model parameters rather than directly manipulating the complex time-domain signal, the system achieves PAPR reduction with reduced processing complexity, effectively creating an optimized copy of the signal through parameter control
Data Source
AI summary
Embodiments of the present disclosure relate to methods, devices and computer readable mediums for a unified nonlinear precoding procedure. The method comprises generating sequences of OFDM symbols on subcarriers on a downlink channel from the network device to a terminal device; determining a processing matrix of the downlink channel at least partially based on characteristics of the downlink channel; obtaining reference values of parameters of a signal generation model; and iteratively performing the following for a number of times: adjusting, based on the sequences of OFDM symbols and the processing matrix, the reference values of the parameters of the signal generation model, the parameters being signals for transmitting via transmitting ports of the network device on the downlink channel to the terminal device; and updating the reference values of the parameters of the signal generation model by clipping the adjusted values of the parameters of the signal generation model.