Abstract:Federated learning (FL) is considered a promising machine learning technique that has attracted increasing attention in recent years. Instead of centralizing data in one location for training a global model, FL allows the model training to occur on user devices, such as smartphones, IoT devices, or local servers, thereby respecting data privacy and security. However, implementing FL in wireless communication faces a significant challenge due to the inherent unpredictability and constant fluctuations in channel characteristics. A key challenge in implementing FL over wireless communication lies in optimizing energy efficiency. This holds significant importance, especially considering user devices with restricted power resources. On the other hand, unmanned aerial vehicle (UAV) technologies present a cost-effective solution owing to flexibility and mobility compared to terrestrial base stations. Consequently, the deployment of UAV communication in FL is viewed as a potential approach to deal with the energy efficiency challenge. In this paper, we address the problem of minimizing the total energy consumption of all user equipment (UE) during the training phase of FL over a UAV communication network. Our proposed system facilitates UE to operate concurrently at the same time and frequency, thereby improving bandwidth utilization efficiently. In this paper, we address the problem of minimizing the total energy consumption during the training phase of FL over a UAV communication network. To deal with the proposed nonconvex problem, we propose a novel alternating optimization approach by dividing the problem into two suboptimal problems. We then develop iterative algorithms based on the inner approximation method, yielding at least one locally optimal solution. The numerical results demonstrate the superiority of the proposed algorithm in solving the proposed problem compared to other benchmark algorithms, particularly in determining the optimal trajectory of the UAVs. In addition, we conduct extensive experiments to evaluate how different parameter settings affect performance after implementing the proposed optimization approaches for deploying FL within the UAV communication system. These analyses yield valuable insights into the comparative effectiveness of the proposed optimization algorithms concerning overall energy consumption reduction.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to minimize the total energy consumption of all user equipments (UEs) during the training phase when implementing federated learning (FL) in unmanned aerial vehicle (UAV) communication networks. Specifically, the paper focuses on the energy - efficiency optimization challenges faced when implementing FL in wireless communication networks, especially when using UAVs as communication nodes. ### Background and Problem Description With the rapid expansion of Internet of Things (IoT) applications and the improvement of the computing and storage capabilities of smart devices, a large amount of data is generated in modern distributed networks every day. Traditional centralized machine - learning methods rely on cloud servers for data storage and training, but this method faces bottlenecks, high costs and latency issues when dealing with large - scale data, especially in real - time applications. In addition, transmitting data from user equipments to third - party cloud locations also raises privacy issues. As a solution, federated learning (FL) allows models to be trained on remote devices without centralizing the original data in one location, thus protecting data privacy and security. However, implementing FL in wireless communication networks still faces multiple challenges, including latency, reliability, scalability and especially energy - efficiency problems. Especially for mobile devices and IoT devices, these devices usually have limited battery capacity, so it is crucial to efficiently manage the energy consumption in local computing and model transmission. ### Research Objectives This paper aims to solve the above problems by designing a UAV communication network that supports FL, in which user equipments operate simultaneously at the same time and frequency to improve bandwidth utilization. The main objectives of the research are: 1. **Minimize Total Energy Consumption**: By optimizing the energy consumption in the computing and communication processes, reduce the total energy consumption of all user equipments during the entire FL training process. 2. **Optimize Resource Allocation**: In particular, determine the optimal flight trajectory of the UAV to further reduce energy consumption. 3. **Propose an Effective Optimization Algorithm**: For the proposed non - convex optimization problem, propose an alternating optimization algorithm, and transform it into a series of convex sub - problems through the inner approximation (IA) method, so as to obtain at least a locally optimal solution. ### Main Contributions - **System Model**: Propose a computational model that supports FL in UAV communication networks, in which all user equipments operate at the same time and frequency. - **Optimization Problem**: Construct an optimization problem aiming to minimize the total energy consumption of all user equipments during the FL training process, covering the energy consumption in both computing and communication. - **Alternating Optimization Algorithm**: By decomposing the original problem into several more tractable non - convex sub - problems and using the inner approximation method to transform them into successive convex sub - problems, propose a new alternating optimization algorithm. - **Numerical Results**: Through simulation results, show the effectiveness of the proposed algorithm in solving the given problem and determining the optimal position of the UAV, indicating that this algorithm significantly reduces the total energy consumption of all user equipments during the FL training process and is superior to other benchmark algorithms. - **Complexity Analysis**: Conduct a detailed complexity analysis of the proposed algorithm to verify its feasibility and acceptable complexity cost. Through these contributions, this research provides an effective method for efficiently implementing federated learning in UAV communication networks, especially in energy - limited IoT devices.

Optimization of Energy Efficiency for Federated Learning over Unmanned Aerial Vehicle Communication Networks

Non-orthogonal Multiple Access Assisted Federated Learning for UAV Swarms: an Approach of Latency Minimization

Optimization algorithm of UAV‐assisted federated learning resources based on wireless energy harvesting

UAV Communications for Sustainable Federated Learning

Wireless Federated Learning over UAV-enabled Integrated Sensing and Communication

When UAV Meets Federated Learning: Latency Minimization via Joint Trajectory Design and Resource Allocation

A Federated Learning Latency Minimization Method for UAV Swarms Aided by Communication Compression and Energy Allocation

Energy-Efficient Dynamic Device Scheduling for Over-the-Air Federated Learning in UAV Swarms

Federated Learning via Unmanned Aerial Vehicle

Joint Training and Resource Allocation Optimization for Federated Learning in UAV Swarm

Federated Learning Via Nonorthogonal Multiple Access for UAV-Assisted Internet of Things

Federated Learning Via Non-Orthogonal Multiple Access for UAV-Assisted Internet of Things

Energy Consumption Minimization for Hybrid Federated Learning and Offloadable Tasks in UAV-Enabled WPCN

Optimizing Data-Driven Federated Learning in UAV Networks

Online Optimization for Over-the-Air Federated Learning with Energy Harvesting

Reinforcement-Learning-Based Optimization on Energy Efficiency in UAV Networks for IoT

UAV-Enabled Federated Learning in Dynamic Environments: Efficiency and Security Trade-off

Joint Trajectory and Communication Optimization for UAV-Assisted Over-The-Air Federated Learning.

Energy Efficient Federated Learning Over Wireless Communication Networks

Energy Efficiency Optimization of UAV-Assisted Wireless Powered Systems for Dependable Data Collections in Internet of Things

Multi-UAV-Assisted Federated Learning for Energy-Aware Distributed Edge Training