ML Digital Pre-Distortion for Power Amplifier Linearity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Power amplifiers in communication systems face non-linearity issues that lead to spectral regrowth, adjacent channel interference, and in-band distortion, resulting in degraded bit-error rate and data throughput, and existing methods for digital pre-distortion are inadequate for high bandwidth and dynamic conditions.
Innovation Solution
The implementation of machine learning-based digital pre-distortion systems that use internal feedback receivers to emulate external feedback signals, allowing for the determination of digital pre-distortion parameters using machine learning models, enabling improved linearity and efficiency of power amplifiers even in field deployment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning models are used to determine digital pre-distortion parameters, then linearity and efficiency of power amplifiers are improved, but device complexity and computational requirements increase
Solution Approach 1:
The machine learning models are trained in advance during manufacturing to learn the inverse characteristics of the power amplifier. This preliminary training phase allows the models to be stored in memory, eliminating the need for complex real-time computation during operation. The models are pre-computed to map input signals to appropriate pre-distortion parameters, reducing operational complexity while maintaining improved linearity.
Solution Approach 2:
Traditional digital pre-distortion systems use complex mathematical computations and iterative algorithms to determine correction parameters. This patent replaces that mechanical/computational system with machine learning models that perform the same function through learned patterns. The ML models substitute complex real-time computation with faster inference operations, reducing device complexity while achieving better linearity performance.
2Device complexity
If digital pre-distortion parameters are determined using traditional methods, then device complexity is kept low, but linearity and spectral purity deteriorate due to non-linearity effects
Solution Approach 1:
The patent changes the approach from fixed, pre-computed pre-distortion parameters to dynamic parameters determined by machine learning models. The ML models continuously adapt parameters based on actual power amplifier behavior and operating conditions. This parameter change enables better compensation of non-linear effects, reducing spectral regrowth and adjacent channel interference while maintaining manageable device complexity through efficient model architectures.
3Ease of operation
If internal feedback receivers are used instead of external feedback receivers, then ease of operation and field deployability are improved, but measurement precision and feedback signal quality deteriorate
Solution Approach 1:
The patent uses machine learning to create a virtual copy of the external feedback receiver's functionality using only the internal feedback receiver. The ML model learns the relationship between internal and external feedback signals during manufacturing, then replicates the external receiver's measurement capabilities using the internal receiver during field operation. This copying approach enables field deployability with internal receivers while maintaining measurement precision equivalent to external receivers.
Solution Approach 2:
The system implements a feedback mechanism where the internal feedback receiver's limitations are compensated through machine learning. The ML models use the internal feedback signal, enhance it through learned corrections, and generate accurate pre-distortion parameters. This feedback-based approach allows the system to overcome the inherent limitations of internal receivers, maintaining measurement precision while enabling ease of operation and field deployment.
Data Source
AI summary
Example embodiments relate to machine learning based digital pre-distortion for power amplifiers. A device may amplify a signal with a power amplifier and transmit the signal. The signal may be received by an internal feedback receiver of the device. The device may further comprise a first machine learning model configured to emulate an external feedback receiver and to generate an emulated feedback signal based on the internal feedback signal. The device may further comprise a second machine learning model configured to determine digital pre-distortion parameters for the power amplifier based on the emulated feedback signal. Apparatuses, methods, and computer programs are disclosed.