A Cognitive Architecture for Shared Autonomy in AUV Operations
Robotics
Summary
The authors created a smart system to help people control underwater robots more safely and easily. They combined a structured knowledge base (an ontology) with advanced language models (LLMs) that understand mission details and help plan and carry out tasks. Their system checks whether a mission is doable, helps plan it, and simulates execution while letting the human operator stay involved. They tested different LLMs and found that GPT-OSS worked best for planning and mission control, while Qwen2.5 was better at understanding mission descriptions in regular language.
Remotely Operated Vehicle (ROV)Unmanned Underwater Vehicle (UUV)OntologyLarge Language Models (LLMs)Situational AwarenessMission PlanningSimulationNatural Language ProcessingHuman-Machine Interaction
Authors
Niamh Ellis, Thi Tran, Ignacio Carlucho, Yvan R. Petillot
Abstract
Operators remain essential to Remotely Operated Vehicle (ROV) operation, yet often suffer from low situational awareness and high workload, both of which negatively affect safety. This paper presents a cognitive architecture consisting of an ontology and multiple Large Language Models (LLMs) to assist the operator at all stages of the mission. Each LLM is grounded with domain-specific information from the ontology and given a simple role to create a system that can support the operator at all stages of an operation. We are aiming to prove that using the two together will allow decisions to be grounded in the relevant domain knowledge, but also benefit from the reasoning capabilities of the LLM. Our framework determines if a mission is possible for a given Unmanned Underwater Vehicle (UUV), performs mission planning, and executes a given mission in simulation. The operator can be involved in planning and execution, ensuring the resulting plan is valid and that the vehicle behaves safely during execution. We compare different LLMs, Llama3, GPT-OSS, and Qwen2.5, to determine which are best suited to the different roles within our framework. We find that GPT-OSS performs best for feasibility assessment, planning, and execution, while Qwen2.5 is best suited to identifying mission types from natural language input.