A Multi-Feature Spectrum Sensing Algorithm Based on GRU-AM

Abstract

In cognitive radio (CR), spectrum sensing aims to detect unused spectrum within a frequency range for exploitation. However, the complexity and diversity of communication environments render this detection task challenging. This letter proposes a method that first extracts multiple features from signals, followed by the integration of Gated Recurrent Units (GRU) and an attention mechanism (AM), aiming to enhance the accuracy and robustness of spectrum sensing. Experimental results demonstrate this approach exhibits favorable sensing performance, particularly enhancing accuracy in low signal-to-noise ratio (SNR) conditions, significantly contributing to improved spectrum utilization efficiency. Additionally, it has low computational complexity.