Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Decentralized Multi-Robot Obstacle Detection and Tracking in a Maritime Scenario

Muhammad Farhan Ahmed, Vincent Frémont

TL;DR

This work tackles robust perception for maritime multi-robot teams faced with specular water, low texture, and limited communications. It presents a decentralized pipeline that fuses YOLOv8+stereo detections with per-object EKF tracking and covariance-intersection fusion, complemented by an information-driven capacitated min-cost flow allocator and hover-view planning. The approach yields improved coverage, localization accuracy, and tracking consistency while keeping bandwidth modest, with explicit mode switching between Surveillance and Tracking to preserve observability. Practical significance lies in enabling scalable, robust maritime monitoring and recovery tasks in real-world, communication-constrained environments.

Abstract

Autonomous aerial-surface robot teams are promising for maritime monitoring. Robust deployment requires reliable perception over reflective water and scalable coordination under limited communication. We present a decentralized multi-robot framework for detecting and tracking floating containers using multiple UAVs cooperating with an autonomous surface vessel. Each UAV performs YOLOv8 and stereo-disparity-based visual detection, then tracks targets with per-object EKFs using uncertainty-aware data association. Compact track summaries are exchanged and fused conservatively via covariance intersection, ensuring consistency under unknown correlations. An information-driven assignment module allocates targets and selects UAV hover viewpoints by trading expected uncertainty reduction against travel effort and safety separation. Simulation results in a maritime scenario demonstrate improved coverage, localization accuracy, and tracking consistency while maintaining modest communication requirements.

Decentralized Multi-Robot Obstacle Detection and Tracking in a Maritime Scenario

TL;DR

This work tackles robust perception for maritime multi-robot teams faced with specular water, low texture, and limited communications. It presents a decentralized pipeline that fuses YOLOv8+stereo detections with per-object EKF tracking and covariance-intersection fusion, complemented by an information-driven capacitated min-cost flow allocator and hover-view planning. The approach yields improved coverage, localization accuracy, and tracking consistency while keeping bandwidth modest, with explicit mode switching between Surveillance and Tracking to preserve observability. Practical significance lies in enabling scalable, robust maritime monitoring and recovery tasks in real-world, communication-constrained environments.

Abstract

Autonomous aerial-surface robot teams are promising for maritime monitoring. Robust deployment requires reliable perception over reflective water and scalable coordination under limited communication. We present a decentralized multi-robot framework for detecting and tracking floating containers using multiple UAVs cooperating with an autonomous surface vessel. Each UAV performs YOLOv8 and stereo-disparity-based visual detection, then tracks targets with per-object EKFs using uncertainty-aware data association. Compact track summaries are exchanged and fused conservatively via covariance intersection, ensuring consistency under unknown correlations. An information-driven assignment module allocates targets and selects UAV hover viewpoints by trading expected uncertainty reduction against travel effort and safety separation. Simulation results in a maritime scenario demonstrate improved coverage, localization accuracy, and tracking consistency while maintaining modest communication requirements.
Paper Structure (29 sections, 20 equations, 10 figures, 1 table, 1 algorithm)

This paper contains 29 sections, 20 equations, 10 figures, 1 table, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORK
  3. Maritime Visual Detection
  4. Multi-Object Tracking
  5. Decentralized Fusion
  6. Assignment and Task Allocation
  7. METHODOLOGY
  8. System Overview and Notation
  9. Coordinate Frames and Transforms
  10. Agent Localization via EKF
  11. YOLOv8 Detection and disparity
  12. YOLOv8 Detection.
  13. Stereo 3D lifting via disparity
  14. Multi-Object Tracking with EKF
  15. Decentralized Fusion via Covariance Intersection
  16. ...and 14 more sections

Figures (10)

  • Figure 1: System-level flow with UAV (green) and Aquabot (blue) ROS nodes.
  • Figure 2: System TFs Overview.
  • Figure 3: Initial pos. of (a) Aquabot, containers and (b) drones.
  • Figure 4: The mission trajectory, Drone1(blue), Drone2(orange) and Drone3(green).
  • Figure 5: Containers (C1--C5): RGB images (top row) and disparity images (bottom row).
  • ...and 5 more figures