SPACE Simulator Documentation

Welcome to the official documentation for SPACE (Swarm Planning And Control Evaluation) Simulator. This simulator is designed to facilitate research and development in swarm robotics and decentralized decision-making. Below are the key features of the SPACE simulator:

Key Features

  • Swarm Robotics Focus: Optimized for swarm robotics research, supporting large-scale simulations with a lightweight design.

  • Behavior Trees: Uses behavior trees to define and manage agent behaviors, allowing for flexible and structured decision-making.

  • Groot2 Integration: Compatible with Groot2 for visualizing and editing the agents’ behavior trees, enhancing ease of use and analysis.

  • Flexible Configuration: Easily configurable via a YAML file, enabling customization without modifying the simulator’s source code.

  • Custom Plugins: Integrates custom decision-making algorithms as plugins, allowing for tailored testing and experimentation.

  • Local Communication and Awareness: Agents communicate locally within specified radii and maintain situational awareness based on their situational awareness ranges.

  • Dynamic Task Generation: Supports the creation of tasks dynamically during the simulation, adapting to evolving scenarios.

  • Algorithm Comparison: Facilitates the comparison of different decision-making algorithms within a consistent simulation environment and supports Monte Carlo tests for statistical analysis.

For detailed installation instructions, please refer to the Installation Guide.

_images/GRAPE_20_agents_200_tasks.gif

Figure 1: GRAPE with 20 agents and 200 tasks.
_images/GRAPE_500_agents_10_tasks.gif

Figure 2: GRAPE with 500 agents and 10 tasks.

Citations

Please cite this work in your papers!