FadeNet: Deep Learning-Based mm-Wave Large-Scale Channel Fading Prediction and its Applications

Ratnam, V. V.; Chen, Hao; Pawar, Sameer; Zhang, Bingwen; Zhang, Charlie Jianzhong; Kim, Young‐Jin; Lee, Soon-Young; Cho, Minsung; Yoon, Sung-Rok

doi:10.1109/access.2020.3048583

Cited by 51 publications

(37 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A similar approach was taken in [125]. Instead of using LSTM, a CNNbased pathloss and shadowing prediction is proposed in [102], where the coordinates of Rx and Tx, the physical environment information, i.e., terrain height, building height, and foliage height, and the visibility condition (LoS/NLoS) are used as input to predict the received power at each position. Instead of using the generalized/specialized ANN, [103] exploits the Random Forest and KNN method to build a prediction model for unmanned aerial vehicle channels, which requires fewer initial parameters, i.e., propagation distance, altitude of Tx and Rx, visibility condition (LoS/NLoS), and link elevation angle.…”

Section: B Ml-enabled Channel Modeling and Predictionmentioning

confidence: 99%

Artificial Intelligence Enabled Radio Propagation for Communications—Part II: Scenario Identification and Channel Modeling

Huang

et al. 2022

IEEE Trans. Antennas Propagat.

View full text Add to dashboard Cite

This two-part paper investigates the application of artificial intelligence (AI) and in particular machine learning (ML) to the study of wireless propagation channels. In Part I, we introduced AI and ML as well as provided a comprehensive survey on ML enabled channel characterization and antennachannel optimization, and in this part (Part II) we review state-ofthe-art literature on scenario identification and channel modeling here. In particular, the key ideas of ML for scenario identification and channel modeling/prediction are presented, and the widely used ML methods for propagation scenario identification and channel modeling and prediction are analyzed and compared. Based on the state-of-art, the future challenges of AI/ML-based channel data processing techniques are given as well.

show abstract

Section: B Ml-enabled Channel Modeling and Predictionmentioning

confidence: 99%

Artificial Intelligence Enabled Radio Propagation for Communications—Part II: Scenario Identification and Channel Modeling

Huang

et al. 2022

IEEE Trans. Antennas Propagat.

View full text Add to dashboard Cite

show abstract

“…It is also possible to include a local terrain map that indicates the altitude of the terrain at each grid point around the sensor. Further kinds of local maps include building indicator maps [30], building height maps [31], [32], or foliage maps [31]. One could even think of using a picture of the surroundings of the sensor as a local map.…”

Section: B) Local Dnn Estimatorsmentioning

confidence: 99%

Radio Map Estimation: A Data-Driven Approach to Spectrum Cartography

Romero,

Kim

2022

Preprint

View full text Add to dashboard Cite

Radio maps can be utilized to characterize a parameter of interest in a communication channel, such as the received signal strength, at every point of a certain geographical region. This article presents an introductory tutorial to radio map estimation, where radio maps are constructed using spatially distributed measurements. After describing the applications of this kind of maps, this article delves into estimation approaches. Starting by simple regression techniques, gradually more sophisticated algorithms are introduced until reaching state-of-the-art estimators. The presentation of this versatile toolkit is accompanied with toy examples to build up intuition and gain insight into the foundations of radio map estimation. As a secondary objective, this article attempts to reconcile the sometimes conflicting terminology in the literature and to connect multiple bodies of literature and sub-communities that have been working separately in this context.

show abstract

“…PL prediction can be considered as a regression problem in ML, where the features extracted from the propagation environment become its input and PL as a continuous variable output. We summarize some of the ML-based approaches for propagation environment modeling and PL prediction [11]- [30] in Table I, highlighting the propagation environment, frequency, key features, training and testing procedures, PL data source, and ML tools such as artificial neural networks (ANNs), random forest (RF), convolutional neural network (CNNs), autoencoder (AE), and support vector regression (SVR). These works showcased the capability of ML-based methods and their potential in improving PL prediction accuracy.…”

Section: A Previous Workmentioning

confidence: 99%

“…Numerous PL prediction models have been established in the literature, which can be classified into three major categories: statistical [2]- [8], deterministic [9], [10], and learningbased models [11]- [30]. Statistical models such as [2] provide a computationally efficient method by fitting particular equations to measurements obtained in different propagation environments [2]- [8].…”

Section: Introductionmentioning

confidence: 99%

Machine Learning-based Urban Canyon Path Loss Prediction using 28 GHz Manhattan Measurements

Gupta,

Du,

Chizhik

et al. 2022

Preprint

View full text Add to dashboard Cite

Large bandwidth at mm-wave is crucial for 5G and beyond but the high path loss (PL) requires highly accurate PL prediction for network planning and optimization. Statistical models with slope-intercept fit fall short in capturing large variations seen in urban canyons, whereas ray-tracing, capable of characterizing site-specific features, faces challenges in describing foliage and street clutter and associated reflection/diffraction ray calculation. Machine learning (ML) is promising but faces three key challenges in PL prediction: 1) insufficient measurement data; 2) lack of extrapolation to new streets; 3) overwhelmingly complex features/models. We propose an ML-based urban canyon PL prediction model based on extensive 28 GHz measurements from Manhattan where street clutters are modeled via a LiDAR point cloud dataset and buildings by a mesh-grid building dataset. We extract expert knowledge-driven street clutter features from the point cloud and aggressively compress 3D-building information using convolutional-autoencoder. Using a new street-bystreet training and testing procedure to improve generalizability, the proposed model using both clutter and building features achieves a prediction error (RMSE) of 4.8 ± 1.1 dB compared to 10.6±4.4 dB and 6.5±2.0 dB for 3GPP LOS and slope-intercept prediction, respectively, where the standard deviation indicates street-by-street variation. By only using four most influential clutter features, RMSE of 5.5 ± 1.1 dB is achieved.

show abstract

FadeNet: Deep Learning-Based mm-Wave Large-Scale Channel Fading Prediction and its Applications

Cited by 51 publications

References 37 publications

Artificial Intelligence Enabled Radio Propagation for Communications—Part II: Scenario Identification and Channel Modeling

Artificial Intelligence Enabled Radio Propagation for Communications—Part II: Scenario Identification and Channel Modeling

Radio Map Estimation: A Data-Driven Approach to Spectrum Cartography

Machine Learning-based Urban Canyon Path Loss Prediction using 28 GHz Manhattan Measurements

Contact Info

Product

Resources

About