Functional Self-Awareness and Metacontrol for Underwater Robot Autonomy

Area of Science:

• Robotics
• Artificial Intelligence
• Autonomous Systems

Background:

• Autonomous systems require robust operation without human intervention, especially in challenging environments like underwater.
• Underwater robots face internal and external disturbances that compromise their operational capabilities and autonomy.
• Long-term autonomy demands perception, action, robustness, and resilience for fault recovery.

Purpose of the Study:

• To analyze operational conditions for autonomous underwater robots, focusing on the UX-1 mine explorer.
• To introduce a knowledge-based self-awareness and metacontrol subsystem for autonomous reconfiguration.
• To enhance the resilience and mission-capability of autonomous underwater robots.

Main Methods:

• Deep modeling of the functional architecture of autonomous robots.
• Ontological reasoning for self-diagnosis and reconfiguration.
• Utilizing functional redundancy for fault tolerance.

Main Results:

• A self-awareness and metacontrol subsystem was developed for autonomous reconfiguration.
• The system enables self-diagnosis and dynamic adjustment of robot subsystems.
• Functional redundancy is leveraged to maintain mission satisfaction during faults.

Conclusions:

• The developed subsystem significantly enhances the resilience of autonomous underwater robots.
• Knowledge-based self-awareness and ontological reasoning are key to robust autonomous operation.
• This approach ensures mission completion even when facing unexpected faults or disturbances.