This work examines the use of unmanned aerial vehicles (UAVs) as mobile anchors to determine the locations of ground users based on the received signal strength (RSS). This is motivated by the significance of user location information in search and rescue operations, post-disaster recovery, and wireless communication systems. By utilizing the Cramér-Rao lower bound (CRLB) as a measure of localization accuracy, we jointly optimize the UAV’s flight trajectory and the users’ transmission scheduling for localization. We propose an iterative solution in which the transmission scheduling and trajectory design subproblems are solved in turn until convergence. We utilize a successive convex approximation (SCA) approach to address the non-convexity of the transmission scheduling subproblem and adopt a gradient descent method to solve the UAV trajectory optimization subproblem. Extensive simulation results verify that our proposed solution outperforms various baselines.

Low Earth orbit (LEO) satellite networks have emerged as an efficient solution to offer communications services worldwide, especially to remote areas uncovered by current terrestrial networks. A network with sufficient satellites deployed can connect any two points on the ground. However, routing for multicast requests is more challenging due to the inherent mobility of LEO satellites. This paper aims to design the shortest routes for multicast requests while reducing the number of link handovers. The formulated problem appears to be an NP-hard problem. We first propose a low-complexity heuristic algorithm, the Minimum Handover Tree (MHT) algorithm, to find a suboptimal solution efficiently. We then further design a Multicast Tree Size Aware Handover (MTSAH) algorithm to improve the MHT algorithm in the tree size. The simulation results demonstrate that our proposed solutions outperform several baselines in the existing work.

This work examines the task assignment, transmission scheduling, and trajectory design of image-surveillance UAVs dispatched to serve on-demand image capture and delivery services to ground users, e.g., from drivers seeking images of traffic jams or security units requesting images of private homes. In each task, the UAVs are required to capture the image of a specified surveillance region and deliver the image to the requesting user before the deadline. The task assignment, transmission scheduling, and trajectory design are jointly determined to maximize the total surveillance area of the completed tasks. We first examine the single-UAV problem and propose an alternating optimization approach that adopts the exact penalty method to promote near-binary solutions and employ successive convex approximation to deal with the nonconvex trajectory optimization. Then, we extend to the multiple-UAV scenario where cross-UAV tasks may require images to be captured and delivered by different UAVs. To enable distributed implementation, we introduce auxiliary deadlines to limit the time available for local tasks and, thus, decouple the joint optimization problem into multiple single-UAV problems that can be solved in parallel following the procedure derived in the previous case. Numerical simulations are provided to demonstrate the effectiveness of the proposed solutions.

In this work, we use unmanned aerial vehicles (UAVs) to collect data from IoT devices on the ground. Each device has an amount of data that can be sent to the UAV during a specific time window. Our objective is to maximize the total profit that the UAV can collect the data from the IoT devices. Since the problem is NP-hard, we propose a heuristic algorithm in three stages to solve the problem. We solve the traveling-salesman problem (TSP) in the first stage to find the UAV's flying trajectory. In the second stage, we propose an algorithm to change the visiting order or remove IoT devices from the flying trajectory if we cannot satisfy their time constraints. In the third stage, we improve the UAV's flying distance established in the second stage. The simulation results show that the proposed algorithms outperform some baselines in terms of total profit and execution time.

This work examines the use of image surveillance UAVs for relay communication between ground users and a remote base station (BS). UAVs take aerial images of the surveillance region and forward them to the BS while serving the uplink transmission demands of ground users. We first consider the single-UAV scenario and jointly determine the UAV’s trajectory, task assignment, user association, and rate allocation by maximizing the sum-log-throughput of the users subject to constraints on the surveillance coverage, image transmission requirements, and relay capacity. The resulting mixed-integer nonlinear programming problem is solved by an inexact block coordinate descent (BCD) algorithm where we inherit ideas from the exact penalty method for mathematical programming with equilibrium constraints to relax the integer constraints and the successive convex approximation approach to address the non-convexity of the trajectory optimization problem. Then, we extend the proposed framework to the case with multiple UAVs that are dispatched to cover a wide surveillance region. The UAVs may complete both relay and surveillance tasks more efficiently through cooperation and proper task allocation for UAVs. A similar BCD algorithm is adopted to solve the problem. Numerical simulations are provided to demonstrate the effectiveness of the proposed scheme over several baseline methods.

With the population of mobile phones, the telecom company has a user's cellular phone signals and its movement trajectory. So, the company can learn about the user's activity habits and predict where the user may go next to meet the need of network resource and service management. In this paper, we propose a mobility prediction framework with Many-to-one Recurrent Neural Network (RNN). First, we extract the place that the user frequently visits (i.e., Points of Interest (POI)) from the user's mobility data through our proposed POI mining method, i.e., acceleration clustering. Then, we propose an adaptive mapping method to map the user's trajectory to a series of POI. Afterward, we use the RNN with Long Short-Term Memory to learn the user's POI series. Finally, we evaluate the prediction performance of the proposed scheme on two different real datasets. The performance shows that the prediction accuracy of our scheme outperforms previous works.