Search

In this study, we explore a three-dimensional trajectory and pick-up design of an unmanned aerial vehicle (UAV) for parcel delivery. In particular, we consider the real-world scenario in which a weight-restricted UAV cannot pick up all parcels within a single route; therefore the parcel delivery must be divided into multiple trajectories while avoiding no-fly zones. We formulate this problem mathematically as the minimization of total delivery time, which jointly optimizes the pick-up indicators, the lengths of the time slots, and the horizontal and vertical trajectories. To address the non-convexity of the formulated mixed-integer nonlinear programming, we employ a successive convex approximation to convert the problem into a convex form concerning optimization variables and utilize a penalty convex-concave procedure to preserve the binary characteristics of the pick-up indicators. Subsequently, we propose an iterative algorithm based on a block decent algorithm to efficiently identify the optimal solution by solving the relaxed convex problem. To address the problem of high computational complexity associated with the optimization-based algorithm, we also present an unsupervised deep learning (DL)-based heuristic algorithm. The simulation results confirm that the proposed schemes achieve considerably shorter delivery times than the baseline schemes in various scenarios. Furthermore, the DL-based scheme requires about 10% longer delivery time than the optimization-based scheme, but it can approximate the UAV strategy with substantially reduced computation time.

This paper considers a multiple unmanned aerial vehicle (UAV)-enabled wireless powered communication network (WPCN). In this WPCN, UAVs broadcast radio frequency (RF) signals to facilitate a wireless power transfer (WPT) during the downlink phase, and ground nodes (GNs) harvest energy from these RF signals and transmit data to their respective UAVs in the uplink phase. To maximize the minimum uplink throughput of GNs, we jointly optimize the scheduling, transmit power of GNs, and trajectory of UAVs, while satisfying the energy neutrality of GNs and the mobility constraints of UAVs. To solve this non-convex optimization problem, we apply a successive convex approximation to divide the original problem into subproblems and make each of them convex for each optimization variable. Subsequently, we propose an iterative algorithm based on a block coordinate descent technique and efficiently find the optimal solution for each convex subproblem. The simulation result reveals that resource allocation and the trajectory of UAVs are strongly influenced by the interference level within the network. Furthermore, the result verifies that the proposed optimization approach significantly outperforms existing baseline schemes by properly coordinating co-channel and cross-link interferences between distinct UAV networks.

This paper presents a novel framework that combines an optimization-based approach with a deep learning (DL)-based approach to devise a cooperative strategy for unmanned aerial vehicles (UAVs) in wireless-based secure communications by leveraging the strengths of both approaches and addressing their respective limitations. We first formulate a joint optimization problem to maximize the minimum achievable secrecy rate while guaranteeing the minimum harvested energy requirement for each ground node by optimizing scheduling, transmit power, and trajectory. To address the difficulty of solving the formulated non-convex mixed-integer nonlinear programming problem, optimization techniques, such as continuous convex approximation and the block coordinate descent algorithm, are used to efficiently find feasible solutions. To tackle the challenges posed by the high computational complexity and initialization sensitivity of the optimization-based approach, we also propose an unsupervised learning-based deep neural network (DNN) structure with a specialized loss function tailored to our goals that allows the DNN to effectively approximate the optimal strategies for UAVs. Finally, we design a pioneering method that integrates the strengths of the two aforementioned approaches, in which the output of the trained DNN serves as the initial values of the optimization variables, and subsequently, optimization techniques are applied to fine-tune these optimization variables, leading to further performance improvements. Through intensive simulations, we confirm that the integrated scheme provides superior performance without constraint violation compared to the DL-based scheme, while guaranteeing faster convergence than the optimization-based scheme.

In this paper, we investigate the optimal resource allocation in unmanned aerial vehicle (UAV)-assisted wireless-powered two-way communications. The communication process considered here consists of two steps. First, the UAV transmits a control signal over wireless links while ground terminals (GTs) receive information and harvest energy simultaneously, with each GT then using the harvested energy to send data to the UAV. We aim to maximize the minimum uplink throughput among GTs while ensuring the minimum requirement of the downlink throughput for each GT by optimizing the time allocation, the transmit power and the trajectory of the UAV along with the energy harvesting ratio of GTs. First, we propose an effective optimization-based approach to address the non-convexity of the formulated problem, which is difficult to solve. Specifically, we apply a successive convex optimization technique to approximate the convex problem for each optimization variable and find the optimal resource management strategy through a block coordinate descent algorithm. To reduce the high computational complexity of the optimization-based approach, we also develop a deep learning (DL)-based approach consisting of an efficient deep neural network framework and a novel training methodology. Simulation results confirm that the proposed schemes show significant performance improvements over existing baseline schemes. We also confirm that the DL-based scheme achieves performance comparable to the optimization-based scheme with a much shorter computation time.