Two-Robot Hazard Detection System: ROS & Gazebo Simulation
Explore a hospital hazard detection system using two autonomous robots built with ROS and Gazebo. Includes navigation, goal mapping, and automated reporting.
GROUP 1 PRESENTATION
Two-Robot Hazard Detection System
Built with ROS & Gazebo
Presented by John Doe
Role: Simulation Setup · Visualisation · Reporting
INTRODUCTION
Project Overview & My Role
Hospital simulation environment built in Gazebo
Two autonomous robots detect and report hazards
My focus: front-end — environment setup, visual design, and reporting
Goal: Make the system output clear, understandable results for the user
PROJECT OVERVIEW
How the System Works
Launch Two Robots
Generate Random Hazard Locations
Autonomous Navigation
Perform Scan at Goal
Generate & Share Report
Both robots operate simultaneously and independently
NAVIGATION CHALLENGE
Setting Valid Goal Positions
Some areas appeared open but were blocked by collision meshes (e.g., nurse station)
Analysed the environment carefully and mapped known valid positions
Result: Robots only navigate to physically reachable locations
HAZARD MARKER DESIGN
Visual Hazard Indicators
Yellow base plate for high visibility
Bold red X shape for clear hazard identification
Consistent design across all hazard types
No collision properties — robots pass through freely
Designed for clarity in the Gazebo simulation environment
ROBOT TRAIL SYSTEM
Visualising Robot Movement
Small coloured dots mark each robot's path
Teal/green trail
Orange trail
Trails update in real-time as robots move
Makes it easy to track coverage and compare paths between both robots
REPORTING SYSTEM
Automated Hazard Reports
Robot ID
identifies which robot made the detection
Location
coordinates of the hazard
Hazard Type
category of the detected hazard
Recommended Action
suggested response
Published to a shared ROS topic — monitor both robots in one place
[ROBOT 1]
Hazard Report
(3.2, 1.8)
Chemical Spill
Evacuate Zone B
[ROBOT 2]
(7.5, 4.1)
Electrical Risk
Isolate Power
CONCLUSION
Key Takeaways & Future Work
Simulation appearance ≠ physical navigability
Areas that look open may be blocked by invisible collision meshes
Add real sensors (e.g. cameras) for actual data-based hazard detection
Moving from simulated results to real sensor-driven detection
Thank you for listening — Questions welcome
- ros
- gazebo
- robotics-simulation
- hazard-detection
- autonomous-navigation
- hospital-automation
- mapping