This research explores how Multimodal Large Language Models can predict human behavior in shared spaces, enabling safer human-robot interaction across various environments.

• Integrates visual and contextual information to predict human actions
• Evaluates system performance across different environments and activity types
• Identifies key challenges in applying MLLMs to real-world prediction scenarios
• Provides insights for improving human behavior prediction accuracy

From a security perspective, this research enables robots to anticipate human actions and respond appropriately, reducing risks in shared spaces and enhancing safety protocols for autonomous systems.

Context-Aware Human Behavior Prediction Using Multimodal Large Language Models: Challenges and Insights