Holistic resource management in UAV-assisted wireless networks

Abstract

Unmanned aerial vehicles (UAVs) are considered as a promising solution to assist terrestrial networks in future wireless networks (i.e., beyond fifth-generation (B5G) and sixth-generation (6G)). The convergence of various technologies requires future wireless networks to provide multiple functionalities, including communication, computing, control, and caching (4C), necessary for applications such as connected robotics and autonomous systems. The majority of existing works consider the developments in 4C individually, which limits the cooperation among 4C for potential gains. UAVs have been recently introduced to supplement mobile edge computing (MEC) in terrestrial networks to reduce network latency by providing mobile resources at the network edge in future wireless networks. However, compared to ground base stations (BSs), the limited resources at the network edge call for holistic management of the resources, which requires joint optimization. We provide a comprehensive review of holistic resource management in UAV-assisted wireless networks. Integrated resource management considers the challenges associated with aerial networks (such as three-dimensional (3D) placement of UAVs, trajectory planning, channel modelling, and backhaul connectivity) and terrestrial networks (such as limited bandwidth, power, and interference). We present architectures (source-UAV-destination and UAV-destination architecture) and 4C in UAV-assisted wireless networks. We then provide a detailed discussion on resource management by categorizing the optimization problems into individual or combinations of two (communication and computation) or three (communication, computation and control). Moreover, solution approaches and performance metrics are discussed and analyzed for different objectives and problem types. We formulate a mathematical framework for holistic resource management to minimize the linear combination of network latency and cost for user association while guaranteeing the offloading, computing, and caching constraints. Binary decision variables are used to allocate offloading and computing resources. Since the decision variables are binary and constraints are linear, the formulated problem is a binary linear programming problem. We propose a heuristic algorithm based on the interior point method by exploiting the optimization structure of the problem to get a sub-optimal solution with less complexity. Simulation results show the effectiveness of the proposed work when compared to the optimal results obtained using branch and bound. Finally, we discuss insight into the potential future research areas to address the challenges of holistic resource management in UAV-assisted wireless networks.