Intelligent and Converged Networks ISSN 2708-6240 2021, 2(1): 16–29 0?/0? pp???–??? DOI: 10.23919/ICN.2021.0004  C All articles included in the journal are copyrighted to the ITU and TUP. This work is available under the CC BY-NC-ND 3.0 IGO license: https://creativecommons.org/licenses/by-nc-nd/3.0/igo/. Radio spectrum awareness using deep learning: Identification of fading channels, signal distortions, medium access control protocols, and cellular systems Yu Zhou  , Hatim Alhazmi, Mohsen H. Alhazmi, Alhussain Almarhabi, Mofadal Alymani, Mingju He, Shengliang Peng, Abdullah Samarkandi, Zikang Sheng, Huaxia Wang, and Yu-Dong Yao Abstract: Radio spectrum awareness, including understanding radio signal activities, is crucial for improving spectrum utilization, detecting security vulnerabilities, and supporting adaptive transmissions. Related tasks include spectrum sensing, identifying systems and terminals, and understanding various protocol layers. In this paper, we investigate various identification and classification tasks related to fading channel parameters, signal distortions, Medium Access Control (MAC) protocols, radio signal types, and cellular systems. Specifically, we utilize deep learning methods in those identification and classification tasks. Performance evaluations demonstrate the effectiveness of deep learning in those radio spectrum awareness tasks. Key words: cellular system; deep learning; signal classification; spectrum awareness; Convolutional Neural Network (CNN) 1 Introduction Spectrum awareness has gained wide attention in wireless communication fields with the development of the Fifth Generation technology standard for broadband cellular networks (5G) in recent years. The scope of spectrum awareness comprises spectrum sensing, identifying radio systems and wireless terminals, and understanding protocol layers, which are essential in  Yu Zhou, Hatim Alhazmi, Mohsen H. Alhazmi, Alhussain Almarhabi, Mofadal Alymani, Mingju He, Abdullah Samarkandi, Zikang Sheng, and Yu-Dong Yao are with the Department of Electrical and Computer Engineering, Stevens Institute of Technology, Hoboken, NJ 07030, USA. E-mail:fyzhou7, halhazm1, malhazmi, aalmarha, malymani, mhe6, asamarka, zsheng2, yyaog@stevens.edu.  Shengliang Peng is with the College of Information Science and Engineering, Huaqiao University, Xiamen 361021, China. E-mail: peng.shengliang@hqu.edu.cn.  Huaxia Wang is with the College of Engineering, Architecture and Technology, Oklahoma State University, Stillwater, OK 74078-1010, USA. E-mail: huaxia.wang@okstate.edu.  To whom correspondence should be addressed. Manuscript received: 2021-01-17; accepted: 2021-01-26 increasing spectrum utilization rate, resolving spectrum scarcity, detecting vulnerable section in the system and smart radio system management, and assisting adaptive transmissions [1--5] , etc. A wide range of publications are about this field. The research in Ref. [6] has explored the multiband spectrum sensing with secondary user hardware limitation, Ref. [7] studied modulation classification, and Ref. [8] was about classification Medium Access Controls (MAC) protocols. Reference [9] even took one step ahead to build an architecture that dynamically updates wireless system parameters based on spectrum awareness to serve applications better. Spectrum awareness can be achieved through various methods. Spectrum awareness could be gained via statistical analysis, Ref. [10] used received signal strength and direction, and Ref. [11] calculated the maximum posterior probability along with marginal particle filtering. With the progress of improving machine learning technologies, spectrum awareness can also be acquired through machine learning and deep learning methods [12--17] . A collision prediction model for multiple frequencies time division multiple access protocol was developed using Convolutional Neural