RME-GAN: A Learning Framework for Radio Map Estimation Based on Conditional Generative Adversarial Network
RME-GAN: A Learning Framework for Radio Map Estimation Based on Conditional Generative Adversarial Network
复制标题
DOI：
10.1109/jiot.2023.3278235
复制
发表时间：
2022-12
影响因子：
10.6
通讯作者：
Songyang Zhang-;Achintha Wijesinghe;Zhi Ding
中科院分区：
文献类型：
--
作者：
Songyang Zhang-;Achintha Wijesinghe;Zhi Ding
室外无线电覆盖图估计是现代物联网（IoT）和蜂窝系统中网络规划和资源管理的重要工具。无线电地图在空间上描述了无线电信号强度分布并提供网络覆盖信息。一个实际问题是从稀疏的无线电强度测量值中估计高分辨率的无线电地图。然而，在许多室外环境中，非均匀分布的测量位置和接入限制对精确的无线电地图估计（RME）和频谱规划构成了挑战。在这项工作中，我们通过整合著名的无线电传播模型并设计一个条件生成对抗网络（cGAN），为RME开发了一个两阶段学习框架。我们首先探索全局信息以提取无线电传播模式。接下来，我们关注局部特征以估计对无线电地图的阴影效应，从而训练和优化cGAN。我们的实验结果证明了基于室外场景中稀疏观测的生成模型所提出的RME框架的有效性。
Outdoor radio coverage map estimation is an important tool for network planning and resource management in modern Internet of Things (IoT) and cellular systems. A radio map spatially describes radio signal strength distribution and provides network coverage information. A practical problem is to estimate fine-resolution radio maps from sparse radio strength measurements. However, nonuniformly positioned measurements and access constraints pose challenges to accurate radio map estimation (RME) and spectrum planning in many outdoor environments. In this work, we develop a two-phase learning framework for RME by integrating well-known radio propagation model and designing a conditional generative adversarial network (cGAN). We first explore global information to extract radio propagation patterns. Next, we focus on the local features to estimate the shadowing effect on radio maps in order to train and optimize the cGAN. Our experimental results demonstrate the efficacy of the proposed framework for RME based on generative models from sparse observations in outdoor scenarios.