Dexterous Robot Hand Development and Intelligent Control
Project Description
This project aims to break through a major bottleneck in robotics: the efficient collection of high-quality data to teach robotic hand dexterous, human-like manipulation skills. We will achieve this by developing a novel multi-degree-of-freedom exoskeleton glove system for high-fidelity force-haptic teleoperation. This project sits at the intersection of robotics, human-computer interaction, and AI. You will help build a system that allows a human operator to intuitively control a five-fingered dexterous robot hand. The key innovation is bidirectional high-DOF force and tactile feedback. This system serves as a powerful tool to efficiently record rich, multi-modal demonstration data (motions + forces) for imitation learning, which will be used to train robots to perform dexterous tasks autonomously.
We seek motivated undergraduates excited about building the future of robotics. A strong foundation in some areas is preferred:
1. Robotics/Hardware Interest: Basic knowledge of kinematics, electronics, or a willingness to work with hardware (3D modelling, manufacturing) is highly valued.
2. Programming: Proficiency in Python is essential. Experience with C++ and/or ROS/ROS 2 is a significant plus.
3. Machine Learning (Bonus): Familiarity with core ML concepts or PyTorch/TensorFlow is beneficial for the imitation learning portion.
We seek motivated undergraduates excited about building the future of robotics. A strong foundation in some areas is preferred:
1. Robotics/Hardware Interest: Basic knowledge of kinematics, electronics, or a willingness to work with hardware (3D modelling, manufacturing) is highly valued.
2. Programming: Proficiency in Python is essential. Experience with C++ and/or ROS/ROS 2 is a significant plus.
3. Machine Learning (Bonus): Familiarity with core ML concepts or PyTorch/TensorFlow is beneficial for the imitation learning portion.
Supervisor
SHEN, Yajing
Quota
2
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP3200
UROP4100
Applicant's Roles
The students will work on a full-stack robotics project, with potential tasks including:
1. System Development: Integrating the exoskeleton glove, robotic hand, and sensors using ROS 2 for real-time control.
2. Force/Tactile Feedback Design: Creating algorithms to generate meaningful force/tactile sensations from robot sensor data.
3. Data Pipeline & ML: Building software to record and process teleoperation data for training imitation learning models.
4. Testing & Evaluation: Benchmarking the system's performance on dexterous manipulation tasks.
1. System Development: Integrating the exoskeleton glove, robotic hand, and sensors using ROS 2 for real-time control.
2. Force/Tactile Feedback Design: Creating algorithms to generate meaningful force/tactile sensations from robot sensor data.
3. Data Pipeline & ML: Building software to record and process teleoperation data for training imitation learning models.
4. Testing & Evaluation: Benchmarking the system's performance on dexterous manipulation tasks.
Applicant's Learning Objectives
Through this project, the students will gain highly sought-after skills:
1. Technical Knowledge: Deepen your understanding of robotics, haptics, real-time systems, and machine learning through direct application.
2. Full-Stack Development Skills: Become proficient in integrating hardware with software (Python, C++, ROS 2) to create a complex, functional system.
3. Research Competence: Develop essential skills in literature review, experimental design, data analysis, and presenting your findings (potentially leading to a co-authored publication).
4. Collaborative Problem-Solving: Learn to tackle open-ended engineering challenges as part of an interdisciplinary team.
1. Technical Knowledge: Deepen your understanding of robotics, haptics, real-time systems, and machine learning through direct application.
2. Full-Stack Development Skills: Become proficient in integrating hardware with software (Python, C++, ROS 2) to create a complex, functional system.
3. Research Competence: Develop essential skills in literature review, experimental design, data analysis, and presenting your findings (potentially leading to a co-authored publication).
4. Collaborative Problem-Solving: Learn to tackle open-ended engineering challenges as part of an interdisciplinary team.
Complexity of the project
Challenging