Optimizing human-robot handovers: the impact of adaptive transport methods

Area of Science:

• Robotics
• Human-Robot Interaction
• Motor Learning

Background:

• Robots increasingly interact physically with humans during object handovers.
• Predictability of robot actions is crucial for smooth human-robot collaboration.
• Adaptive robot behaviors aim to improve handover efficiency and human comfort.

Purpose of the Study:

• To compare an adaptive robot object transport method with a non-adaptive method for human-robot handovers.
• To assess the impact of adaptivity on handover time, human intervention, and subjective ratings.
• To investigate the role of motor learning in human-robot object exchange.

Main Methods:

• A Franka Panda robotic arm with advanced sensors was used for object transport.
• Forty subjects performed 40 handovers each, comparing adaptive and non-adaptive transport.
• Key metrics included physical handover time, early intervention, and subjective trust/safety perceptions.

Main Results:

• Adaptive transport did not significantly increase handover time compared to non-adaptive.
• Non-adaptive transport did not lead to significantly earlier human intervention.
• Subjects reported lower trust and safety perceptions with the adaptive method.

Conclusions:

• While adaptive transport did not hinder efficiency, it negatively impacted user trust.
• Both methods showed significant learning effects, reducing handover time over successive trials.
• Understanding motor learning is vital for developing user-centric robotic systems for industrial applications.