Embodied Robotic locomotion systems for humanoid robots
Project Description
Inspired by human biomechanics, humanoid robots mimic human form and function, allowing them to operate in human-designed environments. Their anthropomorphic structure enables seamless interaction with tools, navigation of stairs, and functioning in spaces meant for people. With advanced locomotion systems, sensory capabilities, and AI, these robots achieve remarkable movement. This project emphasizes the development of humanoid robots as complete systems or as specialized legged subsystems to perform diverse motions using control techniques such as reinforcement learning, imitation learning, and diffusion-based motion generation.
Supervisor
SHI, Ling
Quota
2
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP3200
UROP4100
Applicant's Roles
1. Familiarize themselves with the framework of robotics and relevant literature
2. Data Collection, Robot Modeling, and Simulation:
• Motion data collection
• Gathering and analyzing data to create accurate robot models.
• Utilizing simulation tools to test locomotion and motion strategies.
3. Control Model Design and Enhancement:
• Developing and refining control algorithms for precise and adaptive movements.
• Implementing frameworks to support diverse control policies, including reinforcement and imitation learning.
4. Robot Control Systems Design and Implementation:
• Designing embedded systems and hardware control architectures.
5. Enhancement of Overall Robot Design:
• Improving the mechanical structure for better performance and efficiency.
• Upgrading hardware systems to support advanced functionalities and durability.
2. Data Collection, Robot Modeling, and Simulation:
• Motion data collection
• Gathering and analyzing data to create accurate robot models.
• Utilizing simulation tools to test locomotion and motion strategies.
3. Control Model Design and Enhancement:
• Developing and refining control algorithms for precise and adaptive movements.
• Implementing frameworks to support diverse control policies, including reinforcement and imitation learning.
4. Robot Control Systems Design and Implementation:
• Designing embedded systems and hardware control architectures.
5. Enhancement of Overall Robot Design:
• Improving the mechanical structure for better performance and efficiency.
• Upgrading hardware systems to support advanced functionalities and durability.
Applicant's Learning Objectives
Throughout the project, applicants are expected to learn and apply the necessary systems that underpin humanoid robot development, including mechanical design, control algorithms, and sensory integration. This involves gaining hands-on experience with robotics software frameworks, such as ROS (Robot Operating System), and leveraging simulation tools to test and refine motion strategies. Additionally, applicants will explore advanced AI techniques to optimize control policies and improve robot performance.
Complexity of the project
Challenging