Efficiency-Driven Adaptive Task Planning for Household Robot Based on Hierarchical Item-Environment Cognition.

Area of Science:

• Robotics
• Artificial Intelligence
• Cognitive Science

Background:

• Household robot task planning is complex, with existing methods often failing due to poor understanding of item and environment specifics.
• Current task planning approaches lack the cognitive capacity to handle the nuances of household items and environments effectively.

Purpose of the Study:

• To propose an efficiency-driven adaptive task planning approach for household robots.
• To enhance the cognitive capabilities of robots for executing unfamiliar household services.

Main Methods:

• Developed a multiple semantic attribute-based priori knowledge (MSAPK) framework for item representation.
• Created a long short-term memory (LSTM) based item cognition model for attribute assignment.
• Constructed an environment cognition model linking items to room types for efficient localization.
• Implemented a hierarchical item-environment cognition strategy for adaptive task planning.

Main Results:

• The proposed approach significantly improves the success and efficiency of household task completion by robots.
• Virtual and real-world experiments validated the effectiveness of the hierarchical item-environment cognition strategy.
• Robots demonstrated enhanced flexibility and efficiency in executing diverse household tasks.

Conclusions:

• The hierarchical item-environment cognition approach offers a robust solution for adaptive task planning in household robotics.
• This methodology enhances robot cognitive abilities, leading to more capable and efficient domestic service robots.