Overview of Online Remote Control Architectures for Unmanned Aerial Vehicles

Abstract

This work reviews and compares architectures of online remote control systems for unmanned aerial vehicles (UAVs), with a focus on swarm control. Centralized, decentralized, and hybrid (Edge-Cloud) architectures are analyzed in terms of fault tolerance, scalability, dependence on the communication channel, and ease of integration with a web interface. A hybrid architecture with a microservice server layer is substantiated as a rational choice: the server handles mission planning, task allocation, and operator interaction, while time-critical functions – formation control, local navigation, and obstacle avoidance – are executed onboard, which allows the swarm to keep operating when the server connection is lost. A software and hardware stack is proposed (React, Node.js, Python services, and companion computers running ROS 2), together with a distribution of protocols by communication segment: HTTP/REST and WebSocket for operator interaction, MQTT for server–vehicle exchange, WebRTC for video, and ROS 2/DDS for inter-vehicle communication. The paper also covers two-level navigation integration (global planning on the server and local obstacle avoidance onboard) and security measures for communication channels. Directions for future research include state synchronization between the server and onboard levels and experimental latency evaluation.