Digital twin for mobile robot modeling and simulation

Abstract With the rapid development of mobile robotics, deploying robots in more realistic testing environments to accurately and rapidly validate control algorithms has become a major challenge in the current mainstream. Traditional robot simulation systems are mostly built on a two-dimensional plane, with poor visualization performance and an inability to reflect real testing environments. This poses a significant portion of the research and development costs when validating robot algorithm functionalities. To address the aforementioned issues, this paper proposes a solution based on 3D reconstruction technology to create a twin scene that accurately represents the real simulation environment. Unity3D engine is utilized to develop a 2D LiDAR distance measurement sensor, and a complete control process for mobile robots is built in Matlab/Simulink. This enables hardware-in-the-loop simulation testing. Through experimental validation, it has been demonstrated that prior to conducting full-scale physical simulations, robots can effectively simulate the control logic within the virtual environment. This allows for online adjustments of various parameters in the control algorithms, leading to improved research and development efficiency while reducing potential risks and costs associated with deploying robots in real-world environments.