Multi-Robot Collaborative SLAM

3-agent cooperative mapping with ROS 2, RTAB-Map, and distributed EKF. Real-time occupancy grid merging via multi-master DDS bridge. 40% faster full-environment mapping vs single robot. <5cm inter-agent localization consistency.

Researcher: Krishna Digamarthi | University of Delaware

---

Results

Metric	Value
Mapping speedup	40% faster vs single robot
Inter-agent consistency	<5cm across all agent pairs
Full coverage time	14.5 min (3 robots) vs 24.3 min (1 robot)
Loop closures	31 (vs 12 single robot)
Map merge artifacts	<2% cells

How to Run

ROS_DOMAIN_ID=0 ros2 launch multi_robot robot_0.launch.py
ROS_DOMAIN_ID=1 ros2 launch multi_robot robot_1.launch.py
ROS_DOMAIN_ID=2 ros2 launch multi_robot robot_2.launch.py
ROS_DOMAIN_ID=10 ros2 launch multi_robot map_merger.launch.py

Proof Materials

Type	Location
Mapping comparison	evaluation/mapping_speed_comparison.md
Comm config	coordination/multi_master_config.yaml
Architecture	docs/architecture.md
Proof guide	docs/proof_guide.md

Limitations

• Tested in controlled warehouse environment
• • Map merging assumes overlapping coverage between agents

  • • Multi-master bridge adds ~50ms latency to cross-agent topics

    • Future Work

    • • Task allocation for optimal coverage planning
      • • Dynamic team size (add/remove robots at runtime)
        • • 3D multi-robot SLAM with voxel map merging