Abstract
The performance of existing signal detection methods depends heavily on the amount of prior information acquired by the sensor of interest. Therefore, to improve cognitive radio-based detection in low-signal-to-noise (SNR) environments, we propose a deep learning method-based passive signal detection. A convolution neural network (CNN) and the long short-term memory (LSTM) approach are used to extract the frequency and time domain features of the signal. Our method can detect signal when little to none prior information exists. The simulation experiments verify the probability of detection for our method. The results show that our method is about 4.5-5.5 dB better than a traditional blind detection algorithm under different SNR environments.
Highlights
As 5G communication and Internet of Things (IoT) technologies grow and become more sophisticated, the number of wireless network devices will only increase more dramatically than before
We propose a model based on CLDNN that can be utilized for spectrum sensing
Our experiment found that the training time of convolution neural network (CNN) + long short-term memory (LSTM) networks is 10 times that of CNN networks, so pure CNN networks is efficient if a little performance degradation is acceptable
Summary
As 5G communication and Internet of Things (IoT) technologies grow and become more sophisticated, the number of wireless network devices will only increase more dramatically than before. To improve detection performance in non-cooperative communication, radio signals are artificially generated according to theoretical information, which can be expressed in frequency-domain and time-domain form; noise does not have such information. D. Ke et al.: Blind Detection Techniques for Non-Cooperative Communication Signals Based on DL method is used in this work to extract and integrate theoretical models to distinguish between signal and noise. We propose a method based on CLDNN networks to model additional features and characteristics of a random signal. LSTM has more parameters than general networks; the method will require more training time To overcome this additional computational cost, CNNs can reduce the dimension of input data [23]. We propose a model based on CLDNN that can be utilized for spectrum sensing. Our purpose is to obtain a high Pd and a low Pfa
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
