Multihierarchical interactive task planning: application to mobile robotics

Area of Science:

• Robotics
• Artificial Intelligence
• Human-Computer Interaction

Background:

• Existing human-machine interactive task planning struggles with processing large information volumes and efficient communication.
• A mismatch often exists between symbols optimal for human interaction and those suitable for automatic planning.

Purpose of the Study:

• To propose a solution for human-machine interactive task planning that simultaneously handles large information processing and efficient communication.
• To develop a symbolic model leveraging human-inspired hierarchical knowledge structuring.

Main Methods:

• Formalization of a symbolic environment model using a multihierarchical structure.
• Development of a human-inspired mechanism for knowledge representation and task planning.
• Adaptation of the general planning process for mobile robotic applications.

Main Results:

• The multihierarchical model enables efficient planning even with complex, large-scale information.
• The approach enhances human-machine communication by utilizing user-familiar symbols without sacrificing planning performance.
• Demonstrated suitability through examples and experiments in a mobile robotic context.

Conclusions:

• The proposed multihierarchical symbolic model effectively addresses key challenges in interactive task planning.
• This approach offers a pathway for more efficient and intuitive human-robot collaboration.
• The method shows promise for real-world applications, particularly in robotics.