Deep learning-driven compressed sensing using U-Net for cognitive radio spectrum detection

Key Points

• Enhanced spectrum detection accuracy achieved using the U-Net architecture and deep learning techniques, improving cognitive radio performance.
• Analyzed data shows significant improvements, with over 90% detection accuracy in noisy environments during multiple trials.
• Methodology involved employing a deep learning framework based on compressed sensing to process complex spectrum data efficiently.
• These findings suggest that incorporating deep learning into radio spectrum detection can enhance system capabilities, though real-world applications need further exploration.