What question did this study set out to answer?

To develop an automated reinforcement learning framework that generates multimodal interaction policies for assistive robots collaborating with humans.

July 21, 2025International Journal of Social RoboticsOpen Access

Multimodal Reinforcement Learning for Robots Collaborating with Humans

View Full Paper

Ask AI

Bookmark

Authors

ASAfagh Mehri ShervedaniUniversity of Illinois Chicago SLSiyu LiÅbo Akademi University NMNatawut MonaikulUniversity of Illinois Chicago

Discussion

Member takes

Implication

Experimental evaluation demonstrates effective collaborative task completion and high usability in assistive home robotics, indicating a scalable alternative to handcrafted interaction policies.

Key Points

To develop an automated reinforcement learning framework that generates multimodal interaction policies for assistive robots collaborating with humans.
Trained a reinforcement learning agent using a human-data-driven simulator to coordinate multimodal inputs, including verbal language and physical actions, for home object retrieval.
Implemented a high-level reward function without fine-tuning and enforced state preconditions to accelerate training.
Evaluated the trained interaction manager with human participants in a real-world assistive scenario.
The reinforcement learning policy enabled the robot to accurately infer human intent and coordinate appropriate physical and verbal actions during collaborative tasks.
Real-world user evaluation demonstrated high usability and successful task completion rates without requiring manually authored interaction rules.

Cite This Study

Shervedani et al. (2025) studied this question.

synapsesocial.com/papers/6a12ecbd86514ddae6c0d0fe https://doi.org/10.1007/s12369-025-01287-6

View Full Paper

Ask AI

Bookmark

July 21, 2025International Journal of Social RoboticsOpen Access

Multimodal Reinforcement Learning for Robots Collaborating with Humans

View Full Paper

Ask AI

Bookmark

Authors

ASAfagh Mehri ShervedaniUniversity of Illinois Chicago SLSiyu LiÅbo Akademi University NMNatawut MonaikulUniversity of Illinois Chicago

Discussion

Member takes

Implication

Experimental evaluation demonstrates effective collaborative task completion and high usability in assistive home robotics, indicating a scalable alternative to handcrafted interaction policies.

Key Points

To develop an automated reinforcement learning framework that generates multimodal interaction policies for assistive robots collaborating with humans.
Trained a reinforcement learning agent using a human-data-driven simulator to coordinate multimodal inputs, including verbal language and physical actions, for home object retrieval.
Implemented a high-level reward function without fine-tuning and enforced state preconditions to accelerate training.
Evaluated the trained interaction manager with human participants in a real-world assistive scenario.
The reinforcement learning policy enabled the robot to accurately infer human intent and coordinate appropriate physical and verbal actions during collaborative tasks.
Real-world user evaluation demonstrated high usability and successful task completion rates without requiring manually authored interaction rules.

Cite This Study

Shervedani et al. (2025) studied this question.

synapsesocial.com/papers/6a12ecbd86514ddae6c0d0fe https://doi.org/10.1007/s12369-025-01287-6

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark