The pressure on the radio spectrum increases as more and more IoT devices are deployed, since most of them need to communicate via wireless technology. Spectrum availability and bandwidth are limited, and their shared use poses serious challenges when massive amounts of data need to be transmitted, even after the advent of 5G technology. In search of solutions, there is a growing interest in incorporating cognitive radio technologies into IoT devices [1]. A cognitive radio is a wireless transceiver that can adapt its behavior to the environment, for which the radio automatically selects the best channel in real time. The ultimate goal is the optimal and efficient use of the radio spectrum [2].
The key feature of cognitive radio devices is their spectrum sensing capability: they can detect whether a wireless channel is busy and, if so, recognize the type of modulation used in the channel. This is necessary to detect if a signal from a certain primary user, or even from an interferer, is present in the spectrum. To perform this recognition task, traditionally either matched filters or certain properties of the modulated signals, such as cyclostationarity, have been exploited [3, 4]. Furthermore, deep learning techniques have recently been reported to perform well in the categorization of radio communication signals [5].
However, the above techniques for modulated-signal recognition have high computational complexity and must be tuned or trained ‘off-line’, limiting transceivers to adapt to variations in the environment. In this communication, we will present a new algorithm for spectrum sensing and the categorization of modulated signals that has two main features: (i) it is unsupervised and can handle unforeseen situations in real-time and (ii) it is computationally simple, so that it can operate even with the limited capabilities of common IoT devices. The idea exploits properties of the L1-norm that have been explored in our previous works [6]. Experiments with real and simulated data will demonstrate the effectiveness of the proposed approach.
REFERENCES.
[1] A. Khan, M. Rehmani and A. Rachedi, “Cognitive-radio-based internet of things: Applications, architectures, spectrum related functionalities, and future research directions”, IEEE wireless communications, vol. 24, no. 3, pp. 17-25, 2017.
[2] S. Peyman and S. Haykin, “Fundamentals of cognitive radio”, John Wiley & Sons, 2017.
[3] P. Urriza, E. Rebeiz and D. Cabric, “Multiple antenna cyclostationary spectrum sensing based on the cyclic correlation significance test”, IEEE Journal of Selected Areas in Communications, vol. 31, no. 11, pp. 2185–2195, 2013.
[4] X. Zhang, R. Chai and F. Gao, “Matched filter based spectrum sensing and power level detection for cognitive radio network”, in IEEE Global Conference on Signal and Information Processing (Global SIP), Atlanta, pp. 1267–1270, Dec. 2014.
[5] T. O’Shea, T. Roy and T. Charles Clancy, "Over-the-air deep learning based radio signal classification", IEEE Journal of Selected Topics in Signal Processing, vol. 12, no.1, pp. 168-179, 2018.
[6] J. Camargo, R. Martín-Clemente, S. Hornillo-Mellado and V. Zarzoso, "L1-norm unsupervised Fukunaga-Koontz transform", Signal Processing, vol.182, 2021.
