PMNet: Large-Scale Channel Prediction System for ICASSP 2023 First Pathloss Radio Map Prediction Challenge

Abstract

Large-scale channel prediction, i.e., estimation of the pathloss from geographical/morphological/building maps, is an essential component of wireless network planning. Ray tracing (RT)-based methods have been widely used for many years, but they require significant computational effort that may become prohibitive with the increased network densification and/or use of higher frequencies in B5G/6G systems. In this paper, we propose a data-driven, model-free pathloss map prediction (PMP) method, called PMNet. PMNet uses a supervised learning approach: it is trained on a limited amount of RT (or channel measurement) data and map data. Once trained, PMNet can predict pathloss over location with high accuracy (an RMSE level of 10 −2 ) in a few milliseconds. We further extend PMNet by employing transfer learning (TL). TL allows PMNet to learn a new network scenario quickly (×5.6 faster training) and efficiently (using ×4.5 less data) by transferring knowledge from a pre-trained model, while retaining accuracy. Our results demonstrate that PMNet is a scalable and generalizable ML-based PMP method, showing its potential to be used in several network optimization applications.

UAV-aided Joint Radio Map and 3D Environment Reconstruction using Deep Learning Approaches

“…The proposed model is compared with the following baselines, MLP [19], RadioUNet [22], PMNet [23] and k-nearest neighbor (KNN), which are summarized as follows:…”

“…3) PMNet [23]: The network consists of 5 ResNet-based encoder layers and 7 ResNet-based decoder layers. Each layer applies several parallel atrous convolutions, Maxpooling, and ReLU.…”

“…Deep learning approaches for radio map construction were attempted in [19]- [21], where the features that are fed into the neural network (NN) include building heights, propagation distances, and the environment maps. In [22,23], a convolutional neural network (CNN) based structure was proposed to learn the channel path loss based on 3D city maps. Some studies also proposed to fuse the information from satellite images to learn the characteristics of the propagation environment [24]- [26].…”

DC-Net: A Distant-range Content Interaction Network for Radio Map Construction

“…Subsequently, Zhang et al [6] proposed a two-stage generative adversarial network, referred to as RME-GAN, which improved RM construction accuracy by capturing global radio propagation patterns and learning shadowing effects through geometric and frequency down-sampling. Inspired by the structure of the Deeplabv3 model [7], Lee et al [8] designed PMNet. This ×××, (Corresponding author: Jingjing Yang; Ming Huang.)…”