Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

OceanPlan: Hierarchical Planning and Replanning for Natural Language AUV Piloting in Large-scale Unexplored Ocean Environments

Ruochu Yang, Fumin Zhang, Mengxue Hou

Abstract

We develop a hierarchical LLM-task-motion planning and replanning framework to efficiently ground an abstracted human command into tangible Autonomous Underwater Vehicle (AUV) control through enhanced representations of the world. We also incorporate a holistic replanner to provide real-world feedback with all planners for robust AUV operation. While there has been extensive research in bridging the gap between LLMs and robotic missions, they are unable to guarantee success of AUV applications in the vast and unknown ocean environment. To tackle specific challenges in marine robotics, we design a hierarchical planner to compose executable motion plans, which achieves planning efficiency and solution quality by decomposing long-horizon missions into sub-tasks. At the same time, real-time data stream is obtained by a replanner to address environmental uncertainties during plan execution. Experiments validate that our proposed framework delivers successful AUV performance of long-duration missions through natural language piloting.

OceanPlan: Hierarchical Planning and Replanning for Natural Language AUV Piloting in Large-scale Unexplored Ocean Environments

Abstract

We develop a hierarchical LLM-task-motion planning and replanning framework to efficiently ground an abstracted human command into tangible Autonomous Underwater Vehicle (AUV) control through enhanced representations of the world. We also incorporate a holistic replanner to provide real-world feedback with all planners for robust AUV operation. While there has been extensive research in bridging the gap between LLMs and robotic missions, they are unable to guarantee success of AUV applications in the vast and unknown ocean environment. To tackle specific challenges in marine robotics, we design a hierarchical planner to compose executable motion plans, which achieves planning efficiency and solution quality by decomposing long-horizon missions into sub-tasks. At the same time, real-time data stream is obtained by a replanner to address environmental uncertainties during plan execution. Experiments validate that our proposed framework delivers successful AUV performance of long-duration missions through natural language piloting.
Paper Structure (20 sections, 2 equations, 7 figures, 1 table)

This paper contains 20 sections, 2 equations, 7 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Works
  3. Problem Formulation
  4. LLM Planning
  5. Task Planning
  6. Motion Planning
  7. Methodology
  8. Generalized LLM Planner
  9. HTN Task Planner
  10. DQN Motion Planner
  11. Holistic Replanner
  12. Experiments
  13. Experimental Setup
  14. Mission 1 - Search Aborted Warship
  15. Mission 2 - Search Glider near Coral Reef
  16. ...and 5 more sections

Figures (7)

  • Figure 1: OceanPlan can accomplish AUV missions through natural language commands in the large-scale unexplored ocean.
  • Figure 2: Hierarchical framework of LLM-task-motion planning and replanning.
  • Figure 3: The entire process of EcoMapper searching the aborted warship given an abstracted human command. Each numbered picture corresponds to a specific phase of the process.
  • Figure 4: Quantitative results of ablation studies demonstrate that our method achieves a good balance between efficiency and validity of planning a long-horizon mission given an abstracted command.
  • Figure 5: Preconditions and effects of predefined AUV actions.
  • ...and 2 more figures