This research explores how robots can be taught to understand and execute tasks described in natural human language, enabling more intuitive human-robot collaboration.

Key Innovations:

• Integration of scene understanding, language processing, and policy learning to translate verbal instructions into physical actions
• Systematic framework for language-conditioned manipulation systems
• Addresses critical safety challenges in robot-human interactions

Engineering Impact: This interdisciplinary approach creates more accessible robotic systems that can respond to natural commands rather than requiring specialized programming, potentially revolutionizing factory automation, healthcare assistance, and household robotics.

Bridging Language and Action: A Survey of Language-Conditioned Robot Manipulation