AI Power Amplifier Predistortion for Spectral Regrowth Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Power amplifiers in RF communication systems, particularly in 5G technologies, suffer from nonlinear effects and memory issues leading to spectral regrowth and signal quality degradation, which are exacerbated by high transmission power and limited supply voltage.
Innovation Solution
A method and system using neural networks to model power amplifier circuits and generate pre-distorted signals, optimizing a multi-objective loss function combining time-domain and frequency-domain losses to compensate for distortions, thereby improving signal quality and reducing spectral regrowth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If digital pre-distortion (DPD) is implemented to compensate for power amplifier distortions, then signal quality is improved, but device complexity increases due to the need for mathematical modeling and manual adjustments
Solution Approach 1:
The patent transforms the DPD design from manual mathematical modeling to automated neural network training by changing the optimization parameters. The system uses multi-objective loss functions that automatically adjust model parameters to minimize both time-domain (EVM) and frequency-domain (ACLR) distortions, eliminating the need for manual mathematical model derivation and adjustment.
Solution Approach 2:
The patent replaces the traditional mechanical/mathematical modeling approach with an AI-based neural network system. Instead of manually deriving mathematical models and calculating reverse functions, the system uses neural networks trained with multi-objective loss functions to automatically learn and compensate for power amplifier distortions, significantly reducing integration complexity.
2Reliability
If mathematical modeling is used for DPD design, then compensation effectiveness is improved, but adaptability deteriorates due to the need for manual adjustments for various specification requirements
Solution Approach 1:
The patent creates a universal DPD design framework using neural networks that can adapt to multiple specification requirements simultaneously. The multi-objective loss function incorporates both time-domain (EVM) and frequency-domain (ACLR) specifications, allowing a single model to satisfy multiple communication standards without manual reconfiguration, thereby improving adaptability while maintaining compensation effectiveness.
3Power
If high transmission power is used in 5G edge devices, then communication range is improved, but non-linearity worsens due to limited supply voltage
Solution Approach 1:
The patent applies preliminary action by pre-distorting the signal before it enters the power amplifier. The neural network compensator modifies the input signal to anticipate and counteract the non-linear effects that will occur during amplification, allowing the system to operate at high transmission power levels while maintaining signal quality and minimizing spectral regrowth.
Data Source
AI summary
A compensator compensates for the distortions of a power amplifier circuit. A power amplifier neural network (PAN) is trained to model the power amplifier circuit using pre-determined input and output signal pairs that characterize the power amplifier circuit. Then a compensator is trained to pre-distort a signal received by the PAN. The compensator uses a neural network trained to optimize a loss between a compensator input and a PAN output, and the loss is calculated according to a multi-objective loss function that includes one or more time-domain loss function and one or more frequency-domain loss functions. The trained compensator performs signal compensation to thereby output a pre-distorted signal to the power amplifier circuit.