PA-Aware Neural Constellation Modulation for Phase-Noise-Limited Terminals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing modulation methods such as QAM and APSK suffer significant performance degradation due to phase noise and power amplifier (PA) saturation points, which are not adequately addressed in current adaptive modulation and coding (AMC) techniques.
Innovation Solution
A terminal equipped with a control unit that maps bit sequences to indices and pre-trains a neural network to output modulated symbols, considering power amplifier saturation points, to enhance transmission to a base station.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing modulation methods (QAM, APSK) are used, then transmission can be carried out, but significant performance degradation occurs due to phase noise and PA saturation points
Solution Approach 1:
The patent applies parameter changes by transforming the modulation approach from traditional fixed constellations to neural network-generated adaptive constellations. The neural network learns optimal constellation parameters (amplitude, phase, spacing) that are specifically adapted to counteract phase noise and PA saturation effects, thereby improving transmission performance under these harmful conditions
Solution Approach 2:
The patent substitutes traditional mechanical/mathematical modulation schemes (QAM, APSK) with an intelligent system based on neural networks. Instead of using predetermined modulation constellations, the system employs machine learning models that dynamically generate optimized constellations, replacing conventional signal processing mechanisms with AI-based adaptation
2Speed
If higher frequency and wider bandwidth are used, then data transmission speed increases, but RF oscillator produces increased phase noise
Solution Approach 1:
The patent converts the harmful effect of phase noise into a beneficial factor by training the neural network to recognize and adapt to phase noise characteristics. The network learns to generate constellations that are specifically optimized for noisy conditions, effectively transforming the adverse high-frequency environment into an opportunity for AI-driven optimization
3Reliability
If high-power amplifier is used to gain higher SNR, then signal-to-noise ratio improves, but non-linear distortion is produced
Solution Approach 1:
The patent applies preliminary action by pre-training the neural network offline with extensive data that includes various distortion scenarios. This pre-training equips the network with knowledge of how to compensate for HPA non-linearities before actual transmission occurs, allowing the system to proactively counteract distortion effects rather than reacting to them in real-time
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A terminal includes: a control unit configured to: map bit sequences to indices; output one-hot encoder vectors associated with the indices; and pre-train a neural network (NN) to output modulated symbols in response to the one-hot encoder vectors and a power amplifier (PA) saturation point as inputs; and a transmitting unit configured to carry out a transmission to a base station by using modulated symbols that are selected based on a constellation generated by the pre-trained NN.