DeepSweep: Parallel and Scalable Spectrum Sensing via Convolutional Neural Networks

TL;DR

DeepSweep tackles the need for fast, accurate, and scalable spectrum sensing in modern wireless systems by introducing a parallel, spectrum-sweeping DL-based transceiver. It processes the spectrum in G parallel chunks using a shallow 1D CNN, allowing high-resolution sensing with low inference latency while coexisting with ongoing OFDM demodulation. The approach reduces training and inference times by over 2× and 10× respectively, achieves up to 98% accuracy in locating narrowband interference, and delivers outputs in under 1 ms in over-the-air tests. Validated on an OTA WiFi setup with a wideband 10 MHz channel, the results demonstrate practical viability for real-time spectrum sensing across diverse scenarios and use cases.

Abstract

Spectrum sensing is an essential component of modern wireless networks as it offers a tool to characterize spectrum usage and better utilize it. Deep Learning (DL) has become one of the most used techniques to perform spectrum sensing as they are capable of delivering high accuracy and reliability. However, current techniques suffer from ad-hoc implementations and high complexity, which makes them unsuited for practical deployment on wireless systems where flexibility and fast inference time are necessary to support real-time spectrum sensing. In this paper, we introduce DeepSweep, a novel DL-based transceiver design that allows scalable, accurate, and fast spectrum sensing while maintaining a high level of customizability to adapt its design to a broad range of application scenarios and use cases. DeepSweep is designed to be seamlessly integrated with well-established transceiver designs and leverages shallow convolutional neural network (CNN) to "sweep" the spectrum and process captured IQ samples fast and reliably without interrupting ongoing demodulation and decoding operations. DeepSweep reduces training and inference times by more than 2 times and 10 times respectively, achieves up to 98 percent accuracy in locating spectrum activity, and produces outputs in less than 1 ms, thus showing that DeepSweep can be used for a broad range of spectrum sensing applications and scenarios.

Figures (9)

• Figure 1: A high-level overview of the wireless spectrum, showing how as the technology advances, more of the spectrum is used within our communication.
• Figure 2: The high-level design of DeepSweep, showing how the OFDM RX chain and DeepSweep work in parallel without interruption.
• Figure 3: The architecture of the Spectrum Sensing CNN.
• Figure 4: The experimental setup on the Arena testbed.
• Figure 5: Validation accuracy over the model training for the DeepSweep CNN.