Deep Learning-Based Uplink Sampling Frequency Offset Estimation for Ambient IoT System

Thanks to its unique battery-less feature, ambient Internet of Things (A-IoT) facilitates ultra-low-power communication, positioning it as a key technology in 5th-generation advanced (5G-A) and 6th-generation (6G) communication systems. Due to their low-precision oscillators, A-IoT devices often experience significant sampling frequency offset (SFO), leading to symbol timing errors that accumulate over time and degrade communication performance. Achieving accurate SFO estimation is particularly challenging in traditional spectral analysis-based methods in uplink (UL) transmission, especially for short packet communications. To address this issue, we propose SFO-CLA, a hierarchical feature-sensing deep learning framework that leverages waveform feature extraction for high-precision SFO estimation while ensuring robust performance across diverse conditions. Through link-level simulations (LLS), a remarkable 2.5 dB gain in block error rate (BLER) performance is achieved compared to the traditional method, demonstrating the effectiveness of the proposed approach.

• Publication year

  2025

• Venue

  International Conference on Critical Infrastructure Protection

• Publication date

  2025-11-12

• Fields of study

  Computer Science, Engineering, Environmental Science

• Identifiers
  DOI 10.1145/3784833.3784855
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• An interpretable deep learning framework for intrusion detection in industrial Internet of Things

  2025cited by this paper
• Training Deep Filters for Physical-Layer Frame Synchronization

  2022cited by this paper
• Synchronous Clock Recovery of Photon-Counting Underwater Optical Wireless Communication Based on Deep Learning

  2022cited by this paper
• Internet of Things: Energy boon or bane?

  2019cited by this paper
• Deep Learning-based Frame and Timing Synchronization for End-to-End Communications

  2019cited by this paper
• Data-Driven Deep Learning for Automatic Modulation Recognition in Cognitive Radios

  2019cited by this paper
• Ambient Backscatter Communications: A Contemporary Survey

  2017cited by this paper
• Power of Deep Learning for Channel Estimation and Signal Detection in OFDM Systems

  2017cited by this paper
• An Introduction to Deep Learning for the Physical Layer

  2017cited by this paper
• Attention is All you Need

  2017cited by this paper
• Internet of Things: A Survey on Enabling Technologies, Protocols, and Applications

  2015cited by this paper
• Ambient RF Energy-Harvesting Technologies for Self-Sustainable Standalone Wireless Sensor Platforms

  2014cited by this paper
• ImageNet classification with deep convolutional neural networks

  2012cited by this paper
• 2005 Special Issue: Framewise phoneme classification with bidirectional LSTM and other neural network architectures

  2005cited by this paper
• Long Short-Term Memory

  1997cited by this paper
• Training Stochastic Model Recognition Algorithms as Networks can Lead to Maximum Mutual Information Estimation of Parameters

  1989cited by this paper
• On robust estimation of the location parameter

  1980cited by this paper

• No citing papers are available for this paper.