🤖 Embodied AI for Humanoid/Legged Robotics

Humanoid and legged robots represent the next frontier of embodied AI — machines that can perceive, reason, and physically interact with the world in a human-like manner. This research focuses on developing large AI models and vision-language-action (VLA) frameworks that enable humanoid and legged robots to autonomously navigate, manipulate, and collaborate in complex real-world environments.

Our approach integrates three core pillars:

Foundation Models for Robotic Perception and Control — We develop vision-language-action models that bridge high-level semantic understanding with low-level motor control, enabling robots to interpret natural language instructions and execute complex manipulation and locomotion tasks. Our models leverage large-scale pre-training on multimodal data (vision, language, proprioception) and are fine-tuned for real-world deployment on humanoid platforms.
Bio-Inspired Embodied Intelligence — Drawing inspiration from biological locomotion and sensorimotor systems, we design control architectures that enable robust and adaptive walking, running, climbing, and manipulation on diverse terrains. Our work combines reinforcement learning, model predictive control, and sim-to-real transfer to achieve agile and stable locomotion for legged robots in unstructured environments.
Multimodal Learning for Humanoid Robots — We investigate how robots can learn from multimodal sensory inputs (RGB-D cameras, IMUs, tactile sensors, force/torque sensors) to build rich world models that support whole-body planning and contact-rich manipulation. Our research enables humanoid robots to perform tasks in human-centric environments such as homes, offices, and warehouses.

Embodied AI for Humanoid/Legged Robotics

Embodied AI for Humanoid/Legged Robotics

← Back to all Research Directions

Demo Video

Embodied AI for Humanoid/Legged Robotics — TAS Lab, PolyU

Key Research Directions

Vision-language-action models for robotic task planning and execution
Sim-to-real transfer for robust locomotion on diverse terrains
Whole-body control and contact-rich manipulation for humanoid platforms
Reinforcement learning for adaptive and agile legged locomotion
Multimodal perception (vision, tactile, proprioception) for embodied reasoning
Human-robot interaction and collaborative task execution

Target Applications

Logistics and warehousing: Humanoid robots for package handling, sorting, and delivery
Urban maintenance: Legged robots for inspection and maintenance in complex environments
Healthcare and assistive robotics: Service robots for elderly care and rehabilitation
Search and rescue: Legged robots operating in disaster-stricken environments

Selected Publications (*: Corresponding author)

Integrated Planning and Control on Manifolds: Factor Graph Representation and Toolkit.
Yang, P., Wen, W., Yang, R., Zhang, Y., Hu, J., Chen, Y., Xiao, N., Zhao, J.
IEEE International Conference on Robotics & Automation (ICRA), 2026.
Unified Sufficient Conditions for Exact Convex Relaxation of Nonconvex Optimal Control Problems.
Yang, R., Wen, W., Yang, P., Zhao, Z. and Huang, F.
IEEE Transactions on Aerospace and Electronic Systems, 2025. (IF: 5.7, JCR Q1)
EIRM-RL: Epistemic Integrity Risk Monitoring Inspired Safe Reinforcement Learning for Trustworthy Autonomous Navigation.
Zhang, Y., Wang, Y., Wen, W.
IEEE Internet of Things Journal, 13(2), 3500-3512, 2025. (IF: 8.9, JCR Q1)
Learning Safe, Optimal, Real-Time Flight Interaction with Deep Confidence-enhanced Reachability Guarantee.
Zhang, Y., Wang, Y., Yan, P., Wen, W.
IEEE Transactions on Intelligent Transportation Systems, 2025. (IF: 8.4, JCR Q1, Citations: 2)
Tightly Joined Positioning and Control Model for Unmanned Aerial Vehicles Based on Factor Graph Optimization.
Yang, P., Wen, W.*, Bai, S., Hsu, L.T.
IEEE Transactions on Vehicular Technology, 2025. (IF: 7.1, JCR Q1, Citations: 4)
Online Dynamic Model Calibration for Reliable Control of Quadrotor Based on Factor Graph Optimization.
Yang, P., Wen, W.*, Bai, S., Hu, J.
IEEE Transactions on Aerospace and Electronic Systems, 2025. (IF: 5.7, JCR Q1, Citations: 2)
Safe-Assured Learning-Based Deep SE(3) Motion Joint Planning and Control for UAV Interactions with Dynamic Environments.
Zhang, Y., Wen, W., Yan, P.
IEEE ITSC 2024. (Citations: 4)
Tightly Joining Positioning and Control for Trustworthy Unmanned Aerial Vehicles Based on Factor Graph Optimization in Urban Transportation.
Yang, P., Wen, W.
IEEE ITSC 2023. (Citations: 7)

→ Full publication list

Acknowledgement and Collaborators

This research is supported by The Hong Kong Polytechnic University and industry partners. We collaborate with leading research groups in embodied AI and robotics worldwide.

Projects (1)

Multi-robot Collaborative Operations in Lunar Areas for Regolith Processing

December 03, 2023

Abstract