Advancements in high-frequency communication technologies, including millimeter-wave (mmWave), terahertz (THz), and optical wireless bands, play a crucial role in extending wireless connectivity beyond 5 G. These bands provide ultra-wide bandwidths that enable very high data rates, support dense device deployments, and precise positioning. However, their performance critically depends on maintaining clear Line-of-Sight (LoS) conditions, since Non-Line-of-Sight (NLoS) components suffer from strong attenuation and reflection losses. While mmWave links may provide limited connectivity through NLoS reflections, LoS propagation remains the main factor governing the link budget and reliability of the target applications. In contrast, THz and optical wireless links are almost completely blocked by opaque materials, making LoS assurance essential. This paper tackles the issue of LoS coverage by determining the minimum number and optimal placement of Access Points (APs) required to ensure LoS connectivity in stochastic environments with random obstacles. The environment is modeled as a visibility graph whose nodes represent sub-polygons and edges denote visibility overlaps. Using maximal-clique clustering and maximum-clique packing algorithms, the proposed deterministic framework ensures LoS coverage under all realizations within the modeled stochastic ensemble, achieving up to a 50% reduction in the number of required APs while maintaining over 90% LoS coverage for every realization of obstacle locations.

Optical Wireless Communication (OWC) can complement wireless access services provided over Radio-Frequency (RF) bands by offering abundant additional unlicensed spectrum while enhancing privacy, security, and sustainability. This technology, however, requires Line-of-Sight (LoS) links with no obstructions between transmitter and receiver. In addition, the range of Visible Light Communication (VLC) for wireless access is limited to a few meters when using Light Emitting Diodes (LEDs). Hence, deploying VLC Access Points (APs) to prevent indoor coverage gaps is a challenging task. In this paper, we first characterize the VLC cell range in terms of the three-dimensional (3D) position, orientation, and parameters of the LEDs and the photodetector (PD). Secondly, we propose a graph model to represent the floor plan and, based on it, a clique clustering method that determines the minimum number of APs (and their locations) to ensure LoS coverage. Then, we transform this graph into a connectivity tree, such that the APs can communicate with each other via LoS infrared wireless. Based on our findings, the proposed deployment approach significantly outperforms conventional methods in terms of coverage area and required number of APs, improving the minimum/mean data rate of the VLC network when deployed to give service indoors.

We studied power splitting-based simultaneous wireless information and power transfer (PS-SWIPT) in multiple access channels (MACs), considering the decoding cost and non-linear energy harvesting (EH) constraints at the receiving nodes to study practical limitations of an EH communication system. Under these restrictions, we formulated and analyzed the achievable rate and maximum departure regions in two well-studied scenarios, i.e., a classical PS-SWIPT MAC and a PS-SWIPT MAC with user cooperation. In the classical PS-SWIPT MAC setting, closed-form expressions for the optimal values of the PS factors are derived for two fundamental decoding schemes: simultaneous decoding and successive interference cancellation. In the PS-SWIPT MAC with user cooperation, the joint optimal power allocation for users as well as the optimal PS factor are derived. This reveals that one decoding scheme outperforms the other in the classical PS-SWIPT MAC, depending on the function type of the decoding cost. Finally, it is shown that the cooperation between users can potentially boost the performance of a PS-SWIPT MAC under decoding cost and non-linear EH constraints. Moreover, effects of the decoding cost functions, non-linear EH model and channel quality between the users are studied, and performance characteristics of the system are discussed.

We propose a Partition-based Visibility (PV) graph modeling to find the minimum number of Visible Light Communications (VLC) nodes and their locations for reliable indoor coverage. VLC network offers a low-cost technology on a license-free spectrum to complement the contemporary mobile network services offered on Radio Frequency (RF) bands. However, VLC suffers from propagation limits: Firstly, in presence of opaque obstacles such as walls, doors, and even curtains, strong link blockage is experienced as the power of reflections is much weaker than the power of the Line-of-Sight (LoS) link. Secondly, the received optical power at users drops as the angle of irradiance between a LED lamp and the user increases, imposing a range constraint on the VLC nodes. So, inspired by the Art Gallery Problem, we optimize the number and the locations of VLC nodes by characterizing the PV graph as a dual presentation of the floor plan and a Maximal Clique Clustering algorithm, which is able to solve not only the art gallery problem but also to extend the approach for the range constrained case.

We propose a cell planning method for wireless cellular networks to simultaneously balance the volume shares of the cells and to preserve their regularity. In this framework, we first optimize the base station locations to minimize the coefficient of variation (CoV)-based service demand share metric using the power Voronoi algorithm. Then, we maximize the downlink coverage probability by defining a local regularity optimization problem and solve it through particle swarm, farthest point, average point, and spatial median optimization methods. Finally, we apply the proposed methods on a real network to show how the proposed methods simultaneously minimize both CoV-based service demand share and local regularity metrics.

We study the optimization of base station (BS) localization in cellular networks with non-uniform service demand distribution taking into account a variety of server-side and user-side requirements. To this end, we propose a power Voronoi algorithm (PVA) to obtain uniform traffic volume shares per cell while simultaneously meeting the user-side and server-side requirements of the heterogeneous devices. We describe a numerical method to solve the optimization problem by updating the weights associated to different cells in the PVA. Then, we take into account minimum signal to interference plus noise ratio at all devices as a user-side requirement. It is observed that BSs are concentrated in the areas with condensed demand and high SINR requirements. To show the advantages of the proposed method, the results are compared for different SINR requirement distributions giving insight on the performance of the algorithm.