Wireless communication technologies are evolving rapidly to meet the rising demands of high data traffic applications like autonomous driving and the internet-of-things. However, the reach and quality of these new services stand and fall with the bandwidth for the wireless link.
Thus, the scarcity of the radio spectrum is one of the most challenging issues for future communication networks. One solution is spectrum sensing. This process continuously monitors the densely packed incoming spectrum, identifies currently unused spectral regions and assigns these unused regions to the new link.
Spectrum sensing can be implemented in electronic-based digital signal processors (DSPs). However, the limited bandwidth and intensive computational requirement of these devices leads to an excessive power consumption. This prevents its application for wideband new wireless technologies like 5G, 6G and Terahertz communication systems.
Photonic solutions might be efficient alternatives. Their strengths: ultra-high bandwidth, low loss, and high immunity to electromagnetic interference. Nevertheless, the photonic spectrum sensing techniques proposed so far are severely constrained to low-speed and limited bandwidth applications-not mentioning their demand for complex and bulky optical components. Thus, a simple reconfigurable, wideband, and ultra-fast photonic spectrum sensing, which may outperform existing DSP at 5G and THz frequency bands, has yet to be demonstrated.
The THz Photonics Group of the Technische Universität Braunschweig, Germany led by Prof. Schneider proposed a novel photonic spectrum sensing technique based on stimulated Brillouin scattering (SBS) induced transparency. Related research results are published in Photonics Research, Vol. 9, Issue 8, 2021 (Jaffar Kadum, Ranjan Das, Arijit Misra, Thomas Schneider. Brillouin-scattering-induced transparency enabled reconfigurable sensing of RF signals[J]. Photonics Research, 2021, 9(8): 08001486).
The nonlinear effect of SBS leads to an acoustic wave in the medium which transfers power from one optical wave (called pump) to another one with a lower frequency (probe). Following the conservation of energy, the frequency separation between pump and probe corresponds to that of the acoustic wave. Thus, the pump generates a gain region with a narrow bandwidth for the probe wave.
The gain or loss bandwidth is usually very narrowband (10-30 MHz) but it can be broadened by a modulation of the pump wave. So, if two pump waves have a frequency separation of twice the frequency of the acoustic wave and the pump wave at the lower frequency is modulated, a broadband loss is superimposed with a narrow gain in the frequency region in the middle between the two waves.
By adapting the amplitudes of the gain and loss pumps, the amplification and loss cancel each other out. Thus, the region is transparent. However, due to the strong, narrowband phase change, pulses with bandwidths corresponding to the width of this frequency region will not be amplified or attenuated but, they will be delayed in time.
This is the basic idea behind the proposed spectrum sensing technique and schematically depicted in Figure. The unknown radio frequency (RF) spectrum to be sensed is first modulated to generate time-limited pulses from the otherwise time-unlimited spectrum. These pulses are then modulated on an optical carrier for the exploitation of SBS.
Figure. Schematic diagram of the proposed slow light-based spectrum discriminator.
If the unknown RF spectrum has two frequency components (let's say 1.8 and 5 GHz) the modulated spectrum around the optical carrier consists of two lower and two upper sidebands with a bandwidth corresponding to the inverse of the pulse duration.
If one of these sidebands (like the upper 5GHz one in Figure) falls inside the gain-loss range, the pulse in the time domain (as shown with the blue pulse before the Brillouin medium) will be divided into two, i.e. the fast moving 1.8 GHz component and the slower moving 5 GHz component (the cyan and purple pulses at the output). Thus, the spectrum measurement is transferred to a time measurement, which is much easier.
Jaffar Kadum, Ph.D. student in the group and lead author of this work, says: "In the moment we can just discriminate two frequencies and we need quite long optical fibers as Brillouin medium. However, we are working already on ideas of how to enhance the technique to discriminate a lot of frequencies at once and how to integrate it into a photonic chip".
Compared with other spectrum sensing techniques, the proposed solution is fully reconfigurable and feasible for broad frequency bands and it supports fast, real-time applications. When integrated on a photonic chip, the method might be a promising solution for the spectrum scarcity problem in wireless communication systems.