Abstract
Radio environment map (REM) has emerged as a crucial technology to improve the robustness of intelligent transportation systems (ITS) by enhancing network planning and spectrum resource utilization. To construct a precise REM, optimizing deployment of sensor nodes and increasing spatial interpolation accuracy are two main directions. Given the deployment of sensor nodes, high resolution (HR) spatial interpolation would still bring about huge computing overhead, which is not practical for realtime applications. In order to improve the efficiency and accuracy of REM construction, we propose a super-resolution (SR) based REM construction method, which is composed of Kriging interpolation, dictionary learning and random forest. In our method, both low resolution (LR) and HR REM image sets are generated and trained to obtain a random forest model. With spectrum data from the limited number of sensor nodes, a SR REM can be acquired by the proposed method. Simulation results demonstrate that our method can greatly shorten the construction time of REM while maintaining high accuracy.
Highlights
With the ever-increasing deployment of 5G networks and internet of things, radio spectrum is becoming severely crowded, dynamic and underutilized, bringing about unprecedentedly challenges to manage radio resources instantaneously, especially in vehicular network based intelligent transportation systems (ITS) [1]
We introduce the metrics of our proposed method, which have been widely used in quality evaluations of reconstructed images
1: Pull power spectrum density (PSD) data from sensor nodes at locations Z. 2: Use ordinary Kriging to estimate PSD values for all interpolation points under certain diverse resolutions. 3: Use global PSD data to construct X and Y. 4: Extract patch x from X and y from Y, respectively. 5: Sparse representation for y(Dlr ec) and x(Dhr ec). 6: Principal component analysis. end for 7: Train regression forest and generate random forest model
Summary
With the ever-increasing deployment of 5G networks and internet of things, radio spectrum is becoming severely crowded, dynamic and underutilized, bringing about unprecedentedly challenges to manage radio resources instantaneously, especially in vehicular network based intelligent transportation systems (ITS) [1]. To enhance safety and enable flexible radio resource management in future ITS, radio environment map (REM) is expected to play a vital role in the establishment of robust and high-efficient vehicular networks [2], [3]. To construct a precise REM, optimizing deployment of sensor nodes and enhancing spatial interpolation accuracy are two main directions [5], [6]. Spatial interpolation has to be conducted to estimate the unknown signal power spectrum density (PSD) from limited number of sensors [9]. Among them Kriging interpolation is one of the most frequently applied methods for REM construction due to its high accuracy to fit the irregular patterns of radio spatial environment [16].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
