Deep Completion Autoencoders for Radio Map Estimation

Teganya, Yves; Romero, Daniel

doi:10.48550/arxiv.2005.05964

Cited by 2 publications

(3 citation statements)

References 35 publications

(61 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…and E ξ(i) − ξ(j) 2 can be calculated by (16). Then, take the partial derivatives of L(λ, µ) and set them to zero [13].…”

Section: B Constructing the Residual Shadowing Using Krigingmentioning

confidence: 99%

UAV-aided Radio Map Construction for Wireless Communications and Localization

Liu¹,

Chen²

2021

Preprint

View full text Add to dashboard Cite

Radio maps can be used for source localization, link performance prediction, and wireless relay planning. This paper constructs an air-to-ground radio map to predict the channel gain for each link that connects a ground terminal with a low altitude unmanned aerial vehicle (UAV). The challenge is the insufficiency of measurement samples for a radio map in full dimension, where each data point is 6-dimensional as the transmitter and the receiver each has three spatial degrees of freedom. Classical methods, such as k-nearest neighbor (KNN) and Kriging, may fail for insufficient data. This paper proposes to exploit the propagation property in the geometry of the environment to assist the radio map construction. Specifically, the radio map is built via reconstructing a virtual geometry environment. A multi-class virtual obstacle model embedded in a multi-degree channel model is developed, and a least squares problem is formulated to learn the virtual obstacle map and the model parameters. The paper investigates the partial quasiconvexity of the least squares problem, and based on that, an efficient radio map learning algorithm is developed. In addition, a data driven approach is employed to build a residual shadowing map to further improve the details of the constructed radio map. Our numerical results confirm that the proposed method significantly increases the prediction accuracy compared to a Kriging baseline, and reduces the required measurements by more than a half. When the constructed radio map is applied to received signal strength (RSS) based localization in a dense urban environment, substantial performance improvement is observed where a sub-20-meter accuracy is achieved.

show abstract

“…and E ξ(i) − ξ(j) 2 can be calculated by (16). Then, take the partial derivatives of L(λ, µ) and set them to zero [13].…”

Section: B Constructing the Residual Shadowing Using Krigingmentioning

confidence: 99%

UAV-aided Radio Map Construction for Wireless Communications and Localization

Liu¹,

Chen²

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Other standalone approaches were proposed based on low rank and sparse matrix reconstruction [13] and MMSE estimation [14]. Furthermore, deep learning (DL) has also been considered in literature recently as of the time of this writing [15], [16]. In [15], generative adversarial neural networks are used and the authors in [16] use deep completion auto-encoders.…”

Section: Introductionmentioning

confidence: 99%

“…Furthermore, deep learning (DL) has also been considered in literature recently as of the time of this writing [15], [16]. In [15], generative adversarial neural networks are used and the authors in [16] use deep completion auto-encoders. Like our approach, the cited DL approaches do not rely on user location or statistical channel models and its parameters.…”

Section: Introductionmentioning

confidence: 99%

Spatial Signal Strength Prediction using 3D Maps and Deep Learning

Krijestorac¹,

Hanna²,

Čabrić³

2020

Preprint

View full text Add to dashboard Cite

Machine learning (ML) and artificial neural networks (ANNs) have been successfully applied to simulating complex physics by learning physics models thanks to large data. Inspired by the successes of ANNs in physics modeling, we use deep neural networks (DNNs) to predict the radio signal strength field in an urban environment. Our algorithm relies on samples of signal strength collected across the prediction space and a 3D map of the environment, which enables it to predict the scattering of radio waves through the environment. While already extensive body of research exists in spatial signal strength prediction, our approach differs from most existing approaches in that it does not require the knowledge of the transmitter location, it does not require side channel information such as attenuation and shadowing parameters, and it is the first work, to the best of our knowledge, to use 3D maps to accomplish the task of signal strength prediction. This algorithm is developed for the purpose of placement optimization of a UAV or mobile robot to maximize the signal strength to or from a stationary transceiver but it also holds relevance to dynamic spectrum access networks, cellular coverage design, power control algorithms, etc.

show abstract

Deep Completion Autoencoders for Radio Map Estimation

Cited by 2 publications

References 35 publications

UAV-aided Radio Map Construction for Wireless Communications and Localization

UAV-aided Radio Map Construction for Wireless Communications and Localization

Spatial Signal Strength Prediction using 3D Maps and Deep Learning

Contact Info

Product

Resources

About