Deep Learning Based Fully Progressive Image Super-Resolution Scheme for Channel Estimation in OFDM Systems
Yang Zhang, Jun Hou, Huaijie Liu
Abstract
In this paper, a fully progressive deep learning (DL) channel estimation scheme based on image super-resolution is proposed. Specifically, this scheme takes the channel response at the pilot position as a low resolution image and divides the entire estimation process into multiple stages. At each stage, the image needs to be feature extracted and upsampled to a higher resolution. By gradually increasing the resolution of the image through multiple upsampling stages, the corresponding feature and channel information contained in the image will also be improved. Simulation results demonstrate that this scheme outperforms the conventional and other DL estimation algorithms.
Topics & Concepts
UpsamplingChannel (broadcasting)Artificial intelligenceFeature (linguistics)Computer scienceImage (mathematics)Computer visionImage resolutionFeature detection (computer vision)Feature extractionOrthogonal frequency-division multiplexingProcess (computing)Scheme (mathematics)Pattern recognition (psychology)Image processingMathematicsTelecommunicationsLinguisticsOperating systemMathematical analysisPhilosophyAdvanced Image Processing TechniquesSparse and Compressive Sensing TechniquesBlind Source Separation Techniques