Mobile-Swarm-Navigation

Multi-Robot Autonoumous Semantic Exploration and Navigation

[Code]

The Autonomous Swarm Navigation System leverages advanced AI and robotics frameworks to enable coordinated multi-robot mapping, task allocation, and semantic exploration of dynamic environments. This project was developed to showcase innovative swarm robotics solutions for real-world applications.

Robots performing mapping and depth sensing for semantic understanding of the environment.

System Overview

The system employs a Central Nervous System (powered by a Large Language Model) to:

  • Process textual commands.
  • Coordinate robot tasks through a scheduler.
  • Manage dynamic state updates for swarm efficiency.
Left: Environments used. Right: Robot used.

Unit robots autonomously execute tasks such as:

  • SLAM: Simultaneous Localization and Mapping.
  • Stereo Depth Sensing: For spatial perception.
  • Instance Segmentation: To generate a object map of the environment and add them to a shared dynamic database.

These capabilities enable seamless task navigation, manipulation, and exploration within dynamic environments.

Highlights:

  • Real-time map generation and semantic updates.
Dynamic Semantic Mapping.
  • Coordinated task allocation with minimal redundancy.
Dynamic goal assignment and planning in a swarm of robots.
  • Efficient coverage of unknown environments through collaborative exploration.
Collaborative Exploration.

Future Work

  • Integration with Robotic Arms: Increase of Task Space with added manipulation tasks.
  • Simultaneous Exploration and Navigation: Merging of the 2 phases of Exploration and Navigation.