51³Ô¹Ï

Zhou Zhang

Go back to mentorszhang 

Mechanical Engineering Technology

Project: Digital Twins for Assured Safety of Collaborative Robots Driven by Generative AI

Collaborative robots (cobots) are becoming essential in modern manufacturing, where their ability to work alongside humans enhances productivity and efficiency. However, ensuring safety during their operation remains a major challenge, particularly when Generative AI dynamically drives these systems. This project seeks to explore the integration of digital twins within the Unity game engine as a proactive safety mechanism for cobots.

The proposed research framework will leverage Generative AI to interpret narrative job descriptions, generate commands, and simulate actions in a virtual environment. By incorporating advanced algorithms, such as deep reinforcement learning (DRL), multi-objective evolutionary algorithms (MOEAs), and physics-informed neural networks (PINNs), this project aims to optimize the movement and decision-making of cobots within a digital twin. These algorithms will enable precise motion planning, energy-efficient operations, and real-time adaptability to unforeseen scenarios.

By working on this project, undergraduate students will gain hands-on experience in AI, robotics, and simulation, equipping them with critical skills applicable to modern manufacturing and automation research.

Student’s role

Undergraduate students participating in this project will:

  1. Software Development and Simulation: Students will develop and refine digital twin models using Unity, programming cobot behaviors and interactions within the simulation environment.
  2. Generative AI Integration: Students will experiment with Generative AI to automate cobot command generation, testing how AI-driven task execution compares with traditional programming methods.
  3. Machine Learning and Optimization: Students will implement and test reinforcement learning, evolutionary algorithms, and PINNs to optimize cobot performance within the digital twin.
  4. Data Collection and Analysis: Students will collect real-time data from IoT sensors and integrate it into the digital twin to enhance decision-making.
  5. Safety Validation and AI Explainability: Students will work on AI transparency techniques to ensure safe and understandable robot behavior, particularly in human-robot collaboration settings.
  6. Research Documentation and Presentation: Students will participate in writing research papers and presenting findings at academic conferences and institutional research events.

Student’s criteria

No experience needed 

Outcomes

  1. Anticipate having a functional digital twin of a collaborative robot, demonstrating AI-driven task execution and safety validation. Then, to prepare a paper draft that can be contributed to some conferences or journals, for example, the ASME conference or Electronics Journal.
  2. Hands-on learning experiences for undergraduate students in AI-driven robotics, digital twins, and simulation environments.

Schedule

On-campus meetings will be scheduled based on the agreement with the students to discuss the upcoming week's assignments, ensuring clear goals and structured progress. Additional meetings can be arranged in the evenings or at other convenient times that best fit the students' schedules. This flexibility ensures that students balance research with their academic commitments while making meaningful contributions to the project.

Type of job

Hybrid

Applications