A Comprehensive Simulation Platform for Image-Based Visual Servoing: Unity 3D and ROS 2 Integration with Performance Analysis

• Bruno L. Lima UFMG
• Hugo T. M. Oliveira UFAL
• Glauber R. Leite UFAL
• Andressa M. Oliveira UFRN

Resumo

Testing visual servoing algorithms in real robotic systems can be costly, time-consuming, and often limited by hardware availability and safety constraints. To address these challenges, this work presents a simulation platform that inte-grates Unity 3D with ROS 2 to enable the development and validation of image-based visual servoing (IBVS) strategies in virtual environments. The platform simulates a 6-DOF robotic manipulator with a calibrated RGB-D camera and allows real-time communication through the ROS-TCP-Connector. Feature extraction, control computation, and joint actuation are per-formed using ROS 2 packages, while Unity enables the cre-ation of realistic scenes with configurable lighting and image noise. Three simulation scenarios were designed to assess the system's performance, covering both ideal and visually degraded conditions. Results show that the platform provides a reliable, reproducible, and accessible solution for research and education in visual control of robotic manipulators.

Palavras-chave: Visualization, Three-dimensional displays, Systematics, Education, Noise, Virtual environments, Manipulators, Visual servoing, Performance analysis, Robots, Image-Based Visual Servoing, Robotics Simulation, Unity 3D, ROS 2, Visual Control, Performance Analysis, Manipulability