Three-Dimensional Spectrum Occupancy Measurement Using UAV: Performance Analysis and Algorithm Design

Wei, Zhiqing; Yao, Rubing; Kang, Jie; Chen, Xu; Wu, Huici

doi:10.1109/jsen.2022.3161311

Cited by 9 publications

(8 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Consequently, DL models with a large number of layers and filters, popular in the field of computer vision, have been employed for the interpolation problem. These include statistical PL modeling, orthogonal matching pursuit (OMP), long-short term memory (LSTM), convolutional neural networks (CNNs), CAEs, or a combination between CNNs and generative adversial networks (GANs) [6], [8], [12], [13], [17], [22], [27]. These methods employ convolutionalpooling layer structure, rather than fully-connected ones, to both reduce the computational complexity, and preserve the environmental characteristics, i.e.…”

Section: Related Workmentioning

confidence: 99%

Deep Learning for Reduced Sampling Spatial 3-D REM Reconstruction

Ivanov,

Tonchev,

Poulkov

et al. 2024

IEEE Open J. Commun. Soc.

View full text Add to dashboard Cite

Radio environment maps (REMs) have been established as an important tool in spectrum occupancy characterization toward more efficient coverage planning and design of resource allocation algorithms. The utilization of deep learning (DL) techniques for REM reconstruction, particularly when working with a limited number of samples, has garnered significant research attention owing to its speed and accuracy. This is particularly relevant for spatial three-dimensional REMs, which involve an exponential increase in the number of samples compared to the two-dimensional case. This paper presents a method for determining the optimal sampling grid resolution based on two key criteria, 1) the generated map's similarity to the covariance matrix mean (CMM) of measurements collected in real world (RW) scenarios, and 2) reduced computational complexity. Subsequently, three prominent DL models for REM reconstruction, are evaluated, with the convolutional autoencoder (CAE) achieving the best performance. To enhance its accuracy, a neural network (NN) design approach is introduced, which involves assessing the difference in CMM between the original example and the output for each layer in the NN architecture. This method identifies the layers that introduce the smallest difference and determines their optimal number of filters. Thus, the model's complexity is reduced, and the accuracy is improved. Normalized root mean square error of as little as −35 dB is achieved for a sampling rate of 30%.

show abstract

Section: Related Workmentioning

confidence: 99%

Deep Learning for Reduced Sampling Spatial 3-D REM Reconstruction

Ivanov,

Tonchev,

Poulkov

et al. 2024

IEEE Open J. Commun. Soc.

View full text Add to dashboard Cite

show abstract

“…Alternatively, a similar method was applied to infer the path loss from 2D REMs and floor plan images for different indoor environments using DL for REM estimation [20][21][22]29]. The literature has placed a particular focus on the spatial interpolation of REMs under a constrained number of samples, due to some influential works such as [10,[12][13][14][15]18]. This section describes their main features and limitations, to outline the contributions of this paper.…”

Section: Related Workmentioning

confidence: 99%

“…An important role in the development of such mechanisms is played by the application of a spectrum usage database through spatial REMs, as they allow for the computational relaxation of spectrum-sensing methods and the continual analysis of the spectral efficiency [6,7]. Due to the increasing deployment of dense communication nodes, such as Internet of Things (IoT) smart devices, femto-base stations (BSs), cellular users, and communication UAVs, the REMs can be utilised to aid the wireless technologies' coexistence in a limited range of frequency bands, emergency communications, cellular offloading, spectrum decision and handover coverage optimisation, interference mitigation, and cognitive vehicle-to-everything (V2X) networks [6,[8][9][10]. Thus, REMs represent an important functionality for the development of human-centric cognitive wireless access (HC 2 WA) toward the next-generation of communication networks [11].…”

Section: Introductionmentioning

confidence: 99%

Limited Sampling Spatial Interpolation Evaluation for 3D Radio Environment Mapping

Ivanov,

Tonchev,

Poulkov

et al. 2023

Sensors

View full text Add to dashboard Cite

The increasing densification and diversification of modern and upcoming wireless networks have become an important motivation for the development of agile spectrum sharing. Radio environment maps (REMs) are a basic tool for spectrum utilisation characterisation and adaptive resource allocation, but they need to be estimated through accurate interpolation methods. This work evaluated the performance of two established algorithms for spatial three-dimensional (3D) data collected in two real-world scenarios: indoors, through a mechanical measuring system, and outdoors, through an unmanned aerial vehicle (UAV) for measurement collection. The investigation was undertaken for the complete dataset on two-dimensional (2D) planes of different altitudes and for a subset of limited samples (representing the regions of interest or RoIs), which were combined together to describe the spatial 3D environment. A minimum error of −9.5 dB was achieved for a sampling ratio of 21%. The methods’ performance and the input data were analysed through the resulting Kriging error standard deviation (STD) and the STD of the distances between the measurement and the estimated points. Based on the results, several challenges for the interpolation performance and the analysis of the spatial RoIs are described. They facilitate the future development of 3D spectrum occupancy characterisation in indoor and UAV-based scenarios.

show abstract

“…While there is much research focusing on geolocation, it is only focused on obtaining world coordinates of objects for downstream tasks, such as following a target [7]. Similarly, occupancy networks and other mapping approaches aim at obtaining a map for mapping or scene understanding [8].…”

Section: Related Workmentioning

confidence: 99%

Diminishing Domain Bias by Leveraging Domain Labels in Object Detection on UAVs

Kiefer

Messmer

Zell

2021

2021 20th International Conference on Advanced Robotics (ICAR)

View full text Add to dashboard Cite

This paper introduces a novel approach to video object detection detection and tracking on Unmanned Aerial Vehicles (UAVs). By incorporating metadata, the proposed approach creates a memory map of object locations in actual world coordinates, providing a more robust and interpretable representation of object locations in both, image space and the real world. We use this representation to boost confidences, resulting in improved performance for several temporal computer vision tasks, such as video object detection, short and long-term single and multi-object tracking, and video anomaly detection. These findings confirm the benefits of metadata in enhancing the capabilities of UAVs in the field of temporal computer vision and pave the way for further advancements in this area.

show abstract

Three-Dimensional Spectrum Occupancy Measurement Using UAV: Performance Analysis and Algorithm Design

Cited by 9 publications

References 25 publications

Deep Learning for Reduced Sampling Spatial 3-D REM Reconstruction

Deep Learning for Reduced Sampling Spatial 3-D REM Reconstruction

Limited Sampling Spatial Interpolation Evaluation for 3D Radio Environment Mapping

Diminishing Domain Bias by Leveraging Domain Labels in Object Detection on UAVs

Contact Info

Product

Resources

About