Radio Map Reconstruction With Adaptive Spatial Feature Learning

Radio map reconstruction is a fundamental problem of great relevance in numerous real-world applications, such as network planning and fingerprint localization. Sampling the complete radio map is prohibitively costly in practice and difficult to achieve. Such methods for reconstructing radio maps fr...

Full description

Saved in:
Bibliographic Details
Main Authors: Jie Yang, Wenbin Guo
Format: Article
Language:English
Published: Wiley 2024-01-01
Series:IET Signal Processing
Online Access:http://dx.doi.org/10.1049/2024/7090832
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Radio map reconstruction is a fundamental problem of great relevance in numerous real-world applications, such as network planning and fingerprint localization. Sampling the complete radio map is prohibitively costly in practice and difficult to achieve. Such methods for reconstructing radio maps from a subset of measurements are now gaining additional attention. In this paper, we first explore the spatial features of signals on the radio map and formulate the reconstruction problem as an optimization problem with feature penalties. Then, we propose an iteration algorithm with spatial feature learning to reconstruct signals on the radio map, which improves the reconstruction accuracy by using an adaptive feature dictionary. Numerical examples are given to demonstrate the viability and performance of our method at last.
ISSN:1751-9683