摘要
提出一种基于注意力机制的变分自编码器(Variational Auto-Encoder,VAE)与近端策略优化(Proximal Policy Optimization,PPO)方法,用于太阳光干扰下车辆可见光通信的自适应信道优化。该方法将VAE与PPO创新性地结合,利用VAE精确建模复杂光环境下的干扰特性,通过PPO学习最优发光二极管(Light Emitting Diode,LED)光功率调节策略。通过引入注意力机制,显著提高模型对关键干扰特征的捕捉能力。研究结果表明,所提方法在不同太阳光干扰强度下均优于传统VAE-PPO方法,为智能交通系统中的车辆可见光通信提供了可靠的解决方案。
A Variational Auto-Encoder(VAE)and Proximal Policy Optimization(PPO)method based on attention mechanism are proposed for adaptive channel optimization of vehicle visible light communication under sunlight interference.This method innovatively combines VAE with PPO,accurately models the interference characteristics in complex light environment with VAE,and learns the optimal Light Emitting Diode(LED)light power adjustment strategy through PPO.By introducing attention mechanism,the model’s ability to capture key interference features is significantly improved.The research results show that the proposed method is superior to the traditional VAE-PPO method under different solar interference intensities,which provides a reliable solution for vehicle visible light communication in intelligent transportation system.
作者
牛安睿
NIU Anrui(Beijing University of Posts and Telecommunications,Beijing 100876,China)
出处
《通信电源技术》
2025年第12期25-27,共3页
Telecom Power Technology
关键词
可见光通信
太阳光干扰
变分自编码器(VAE)
自适应信道优化
visible light communication
solar interference
Variational Auto-Encoder(WAE)
adaptive channel optimization
作者简介
牛安睿(2000-),男,陕西西安人,硕士研究生,主要研究方向为6G无线通信、可见光通信。
