Socially Aware Robot Obstacle Avoidance Considering Human Intention and Preferences.

Area of Science:

• Robotics
• Artificial Intelligence
• Human-Robot Interaction

Background:

• Safe human-robot navigation requires accurate prediction of human motion, intention, and preferences.
• Existing multi-agent systems often lack social awareness and robustness to human behavioral uncertainty.

Purpose of the Study:

• To develop a novel framework for socially aware robot obstacle avoidance.
• To create a discretized human model for robotics decision-making algorithms.
• To enhance robot navigation capabilities in human-populated environments.

Main Methods:

• Consolidation of human motion, intention, and preference studies into a discretized human model.
• Development of Cooperative Markov Decision Process (Co-MDP) for socially aware obstacle avoidance.
• Simulation-based verification and human-in-the-loop experiments for anthropomorphic behavior assessment.

Main Results:

• Co-MDP effectively approximates rational human behavior and intention.
• The framework generates socially aware obstacle avoidance behaviors.
• Human participants could not distinguish Co-MDP controlled agents from human operators in experiments.

Conclusions:

• Co-MDP is a feasible and robust algorithm for socially aware robot obstacle avoidance.
• The developed discretized human model enhances robot decision-making.
• Experimental results confirm the anthropomorphic and acceptable nature of Co-MDP for human-robot interaction.