Session: 02-05-01: IoT and Digital Twins

Paper Number: 166650

Building a Digital Twin for Mobile Robots: a Step-By-Step Development Guide 

This paper provides a structured, step-by-step approach to designing and developing a digital twin for mobile robots using Blender 3.0. Digital twins serve as powerful tools for real-time simulation, motion optimization, and performance validation, bridging the gap between virtual modeling and physical implementation. By following this guide, readers can systematically create a functional digital twin that enables seamless data exchange and motion synchronization between the virtual and physical robot. Additionally, machine vision integration enhances the robot’s perception capabilities, particularly in detecting human presence and interacting with its environment.

The development process begins with designing and modeling the robot—whether a hexapod or quadruped—using any CAD software. The model is then imported into Blender for digital twin creation. To facilitate real-time control and synchronization, the Marionette plugin for Blender, developed by knee-Koh, is used as a communication interface, allowing users to animate and coordinate the digital twin’s movements with the physical robot.

This guide explores multiple approaches to generating stable and efficient walking gaits. Readers will learn a traditional rule-based method, where limbs move sequentially with defined offsets to maintain stability, as well as an AI-driven reinforcement learning (RL) approach. By leveraging Blender’s Python scripting capabilities, users can implement an RL algorithm that iteratively optimizes gait performance using a reward-based system. This method enables adaptive gait refinement for improved efficiency and robustness in real-world scenarios.

Once a working gait is established, the physical robot—controlled by an ESP-32 microprocessor—is connected to Blender through serial communication. To ensure accurate motion replication, the inverse kinematics (IK) system is initially switched to forward kinematics (FK) for manual motor calibration. The guide provides detailed instructions on adjusting joint limits, aligning digital and physical motion, and re-enabling IK for continuous gait testing and refinement.

A key advantage of this system is its capability to enhance remote operation in hazardous or industrial environments. Equipped with a camera, the robot can transmit real-time video feeds while being remotely controlled through the digital twin interface. Unlike traditional remote-controlled robots, which rely solely on user inputs, the digital twin provides precise control by allowing operators to visualize the exact joint configurations of the physical robot on a computer screen. This feature enables more accurate maneuvering and reduces the risk of operator errors, making it ideal for industrial monitoring, search-and-rescue missions, and exploration in hazardous areas.

Beyond manual control, the digital twin also supports autonomous navigation on predefined paths within structured environments. By integrating sensor feedback and predefined movement trajectories, the system can facilitate semi-autonomous or fully autonomous navigation, making the robot more versatile for industrial and research applications. The validation process compares the digital twin’s simulated motion with real-world behavior, ensuring that the virtual model accurately represents the physical robot’s joint movements, gait dynamics, and overall performance.

Future enhancements include integrating the digital twin with industrial applications for real-time monitoring and expanding the framework to inspect additional robotic systems. By following this structured development guide, researchers, educators, and robotics enthusiasts can build and optimize digital twins for mobile robots, fostering a deeper understanding of simulation-based robotics development.

Presenting Author: Vedang Chauhan California State University Channel Islands

Presenting Author Biography: Associate Professor in Mechatronics Engineering

Authors:

Charles Przechocki Western New England University
Vedang Chauhan California State University Channel Islands