Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SkyRover: A Modular Simulator for Cross-Domain Pathfinding

Wenhui Ma, Wenhao Li, Bo Jin, Changhong Lu, Xiangfeng Wang

TL;DR

SkyRover addresses the lack of cross-domain UAV–AGV simulators by providing a unified, modular 3D MAPF environment that supports realistic dynamics and multiple solver paradigms. It integrates five interconnected modules—Sim World Zoo, 3D Grid Generator, Unified Algorithm Wrapper, Plan Executor, and System Interface—enabling rapid testing of both classical search and learning-based MAPF methods in UAV-AGV coordination tasks. The framework is demonstrated in warehouse and park scenarios, comparing 3D-A*, 3D-CBS, and a curriculum-trained 3D DCC model, and it supports hardware-in-the-loop controllers like PX4 and Navigation2. Collectively, SkyRover enables rapid prototyping, benchmarking, and scalable experimentation for UAV–AGV cooperation in logistics, surveillance, and inspection tasks.

Abstract

Unmanned Aerial Vehicles (UAVs) and Automated Guided Vehicles (AGVs) increasingly collaborate in logistics, surveillance, inspection tasks and etc. However, existing simulators often focus on a single domain, limiting cross-domain study. This paper presents the SkyRover, a modular simulator for UAV-AGV multi-agent pathfinding (MAPF). SkyRover supports realistic agent dynamics, configurable 3D environments, and convenient APIs for external solvers and learning methods. By unifying ground and aerial operations, it facilitates cross-domain algorithm design, testing, and benchmarking. Experiments highlight SkyRover's capacity for efficient pathfinding and high-fidelity simulations in UAV-AGV coordination. Project is available at https://sites.google.com/view/mapf3d/home.

SkyRover: A Modular Simulator for Cross-Domain Pathfinding

TL;DR

SkyRover addresses the lack of cross-domain UAV–AGV simulators by providing a unified, modular 3D MAPF environment that supports realistic dynamics and multiple solver paradigms. It integrates five interconnected modules—Sim World Zoo, 3D Grid Generator, Unified Algorithm Wrapper, Plan Executor, and System Interface—enabling rapid testing of both classical search and learning-based MAPF methods in UAV-AGV coordination tasks. The framework is demonstrated in warehouse and park scenarios, comparing 3D-A*, 3D-CBS, and a curriculum-trained 3D DCC model, and it supports hardware-in-the-loop controllers like PX4 and Navigation2. Collectively, SkyRover enables rapid prototyping, benchmarking, and scalable experimentation for UAV–AGV cooperation in logistics, surveillance, and inspection tasks.

Abstract

Unmanned Aerial Vehicles (UAVs) and Automated Guided Vehicles (AGVs) increasingly collaborate in logistics, surveillance, inspection tasks and etc. However, existing simulators often focus on a single domain, limiting cross-domain study. This paper presents the SkyRover, a modular simulator for UAV-AGV multi-agent pathfinding (MAPF). SkyRover supports realistic agent dynamics, configurable 3D environments, and convenient APIs for external solvers and learning methods. By unifying ground and aerial operations, it facilitates cross-domain algorithm design, testing, and benchmarking. Experiments highlight SkyRover's capacity for efficient pathfinding and high-fidelity simulations in UAV-AGV coordination. Project is available at https://sites.google.com/view/mapf3d/home.

Paper Structure

This paper contains 16 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. SkyRover Simulator
  4. Sim World Zoo
  5. 3D Grid Generator
  6. Unified Algorithm Wrapper
  7. Plan Executor
  8. System Interface
  9. Experiment
  10. Environment Setup
  11. 3D MAPF Examples
  12. Motion Control Integration
  13. Preliminary Results and Comparison
  14. Conclusion and Future Directions
  15. Limitations
  16. ...and 1 more sections

Figures (2)

  • Figure 1: Main Architecture. SkyRover comprises multiple modules to support cross-domain MAPF.
  • Figure 2: (a) The warehouse Gazebo world, featuring multiple Holybro X500 drones and delivery AGVs; (b) The park scenario, offering more open space for UAV operations; (c) The 3D occupancy grid in RViz. Each dark cell has point cloud data and is thus considered an obstacle; (d) Example of integrating SkyRover with hardware-oriented controllers. PX4 executes drone flight commands and Navigation2 governs the TurtleBot.