AquaChat: An LLM-Guided ROV Framework for Adaptive Inspection of Aquaculture Net Pens

2025-07-19

AI Summary

Conventional aquaculture cage detection relies on pre-programmed scripts or manual remote operation, limiting adaptability to dynamic underwater environments and heterogeneous user requirements. Method: This paper proposes the first large language model (LLM)-driven remotely operated vehicle (ROV) framework tailored for underwater aquaculture. It integrates hierarchical task planning, real-time vision–motion feedback, and high-precision execution control to enable closed-loop, end-to-end autonomous operation—from natural language instruction interpretation and dynamic task re-planning to physical execution. Contribution/Results: By pioneering LLM integration into underwater aquaculture inspection, the framework significantly enhances responsiveness under complex hydrodynamic conditions, low visibility, and variable user commands, while improving detection accuracy and operational efficiency. Extensive experiments in both simulated and real-world aquaculture environments demonstrate substantial performance gains, validating the feasibility and practical potential of language-driven AI for intelligent marine equipment.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Planning, Routing, and Scheduling: Planning with Language Models
• Computer Vision: Language and Vision

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Inspection of aquaculture net pens is essential for maintaining the structural integrity, biosecurity, and operational efficiency of fish farming systems. Traditional inspection approaches rely on pre-programmed missions or manual control, offering limited adaptability to dynamic underwater conditions and user-specific demands. In this study, we propose AquaChat, a novel Remotely Operated Vehicle (ROV) framework that integrates Large Language Models (LLMs) for intelligent and adaptive net pen inspection. The system features a multi-layered architecture: (1) a high-level planning layer that interprets natural language user commands using an LLM to generate symbolic task plans; (2) a mid-level task manager that translates plans into ROV control sequences; and (3) a low-level motion control layer that executes navigation and inspection tasks with precision. Real-time feedback and event-triggered replanning enhance robustness in challenging aquaculture environments. The framework is validated through experiments in both simulated and controlled aquatic environments representative of aquaculture net pens. Results demonstrate improved task flexibility, inspection accuracy, and operational efficiency. AquaChat illustrates the potential of integrating language-based AI with marine robotics to enable intelligent, user-interactive inspection systems for sustainable aquaculture operations.

Problem

• Enables adaptive inspection of aquaculture net pens using LLMs
• Overcomes limitations of pre-programmed or manual ROV control
• Integrates natural language commands for intelligent underwater inspection

Innovation

• LLM-guided ROV for adaptive aquaculture inspection
• Multi-layered architecture with natural language processing
• Real-time feedback and event-triggered replanning