Interest in optical wireless communications (OWC) as a possible complement to RF technology has increased recently. However, the propagation of optical beams through the atmosphere distorts the beam's amplitude and phase, resulting in information loss and significant noise. Arising due to a combination of multiple absorption and scattering events, the distortion of the beam makes optical communication difficult. In some cases, the orbital angular momentum of light (OAM), along with various deep learning algorithms (DL), could be helpful to mitigate the problem and provide high-capacity optical communication links. In this work, we propagate Laguerre-Gaussian (LG) beams with different topological charges ( l ) under diffuse and turbulent conditions and develop a deep-learning classification network to characterize the ‘ l ’ of the LG beam. The proposed method is later implemented using a laboratory setup demonstrating communication on the optical table. The results show that the proposed algorithm can identify the modes of the LG beam with high accuracy even when the optical beam propagates through highly turbulent and scattering media. To demonstrate the robustness of the proposed OWC system, small grayscale images are transmitted over the communication channel. A bit error rate (BER) of only $2.3 \times {10}^{ - 4}$ and $9.7 \times {10}^{ - 4}$ for the tabletop experiment and the simulations, respectively. The demonstrated low BER in the proposed OWC system suggests promising applications for secure and reliable data transmission in adverse atmospheric conditions, highlighting the potential of this method in advancing optical wireless communication technologies.

The next generation of satellite constellations would include LEO satellites with communication link above 100 GHz in order to achieve higher efficiency and capacity of the satellite networks due to large available bandwidth. The LEO satellites require tracking by the ground station receiver, however due to the complexity, inefficiency, and nonlinearity of the phase shifters at THz frequencies, the conventional concept of scanning phased array antenna is less recommended. In this work, we derive an optimization algorithm for an electronically switchable subarrays of phased array at the ground station that maximize the SNR depending on the range of elevation angles of satellite. The algorithm optimizes the SNR by adapting the inter-element spacing between array elements considering the impact of molecular absorption loss and noise, thermal noise from the sky and the ground. In comparison to uniform phased array design with equal spacing, the SNR can be improved for a typical scenario by about 3 dB and thus improving the link performance. To the best knowledge of the authors, this LEO satellite communication algorithm at the sub-THz frequencies is for the first time to be published.

Diffuse optical tomography using deep learning is an emerging technology that has found impressive medical diagnostic applications. However, creating an optical imaging system that uses visible and near-infrared (NIR) light is not straightforward due to photon absorption and multi-scattering by tissues. The high distortion levels caused due to these effects make the image reconstruction incredibly challenging. To overcome these challenges, various techniques have been proposed in the past, with varying success. One of the most successful techniques is the application of deep learning algorithms in diffuse optical tomography. This article discusses the current state-of-the-art diffuse optical tomography systems and comprehensively reviews the deep learning algorithms used in image reconstruction. This article attempts to provide researchers with the necessary background and tools to implement deep learning methods to solve diffuse optical tomography.

The recent expansion of networks of low-earth orbit (LEO) satellites such as Starlink, OneWeb, and Telesat and the evolution of communication systems toward B5G and 6G with densely interconnected devices could generate opportunities for various cyber attacks. As the satellite network offers many crucial services to the public and governmental organizations, cyberattacks pose severe risks to the communication infrastructure. In this study, we propose a random routing algorithm to prevent distributed denial-of-service (DDoS) attacks on an LEO satellite constellation network. The routing algorithm utilizes the classical algorithms, i.e., k-DG, k-DS, k-SP, and k-LO, by introducing randomness and selecting one with weighted probability distribution to increase the uncertainty in the algorithm. The study shows that the proposed random routing algorithm improves the average and median cost of the attacker against DDoS attacks while maintaining the functionality of the network. The algorithm is optimized by formulating a Bayesian optimization problem. In addition to providing an additional level of uncertainty in the routing, there is an improvement of $1.71\%$ in the average cost and $2.05\%$ in the median cost in a typical scenario. The algorithm causes the network to be robust to cyber attacks against LEO Satellite Networks (LSNs), however, similar to any other defensive measures, it reduces the network’s goodput.

With the success and commercialization of 5G, 3GPP has started working toward the sixth generation of communication systems. While 5G explored the concept of non-terrestrial networks like satellites and unmanned aerial vehicles working alongside terrestrial networks, 6G is expected to take this integration a step further, aiming to achieve a more coherent network where satellites and terrestrial infrastructure work together seamlessly. However, the complexity and uniqueness of such networks create numerous attack surfaces that make them vulnerable to cyberattacks. The solution to such cyberattacks can be addressed by encryption and other upper-layer authentication methods. However, with the move to higher-frequency bands, such encryption techniques are difficult to scale for low-latency networks. In addition, the recent progress in quantum computing will make networks more vulnerable. To address such challenges, physical layer security (PLS) is proposed as a secure and quantum-resistant way to implement security by taking advantage of the physics of the channel and transceiver. This article reviews the latest trends and progress in PLS in integrated satellite–terrestrial networks (ISTNs) from a signal processing perspective. This work provides a comprehensive survey of the state-of-the-art research conducted, challenges, and future directions in the PLS of ISTNs.

The integration of satellite and terrestrial networks would generate more opportunities for various kinds of cyber attacks, such as spoofing. In this study, we propose a novel signal-processing algorithm for improving the physical layer security of integrated satellite-terrestrial networks (ISTN). The proposed algorithm is based on utilizing the atmospheric signature associated with the received signal at the ground station that causes rapid variations in the signal. We leverage the characteristics of spatial time-frequency signature caused by atmospheric channel associated with incoming signals at the ground station from different satellites to differentiate among the sources from which the signal has been transmitted. The various parameters extracted from the rapidly varying signal component are then provided as inputs to the machine learning algorithm for detecting the legitimate satellite. The algorithm is validated against the measured data obtained at a site located in Israel. The results shows very low false alarm (< 2%) and missed detection rate (< 1%) while the authentication rate is quite high (> 97%).