Case Study
Bringing robotic inspection to industrial training while skipping weeks of setup.
Robots are central to the future of industrial inspection: keeping people out of hazardous zones, delivering consistent readings across hundreds of data points, and preparing students for Industry 4.0. DarbyTech needed to put that into practice.
The Challenge
The requirement was straightforward: an off-the-shelf solution where robot setup and inspection planning doesn't require specialist programming skills.
OLO provided robot control, navigation, and a data layer in the browser. Turning a hardware investment into a repeatable inspection routine, not a one-off demo.
The Approach
The Robot and Payload
We selected a DEEP Robotics M20 Pro wheeled quadruped. Its robustness, long runtime, and agility prompted the choice.
Thermal imaging and RGB PTZ cameras were added alongside the inbuilt 2D LiDAR.
To communicate with the platform, the robot was running the OLO appliance Docker container installed on its onboard computer.
The Environment
The DarbyTech post combustion carbon capture training rig is installed at Redcar and Cleveland College in the UK.
This mimics industrial environments and is the perfect test-bed for training students and industry in future technologies.
How It Worked
Planning in simulation
- 1.A 3D model of the environment to be inspected is imported into the OLO Gazebo simulator.
- 2.The robot to be used for the inspection is then spawned in the simulation along with the required sensors.
- 3.A SLAM inspection map is created in the simulation and waypoints positioned. The actions to perform at each waypoint are defined, and the robot is set to navigate a trial inspection.
- 4.The dashboard is set up with panels playing back each data feed alongside the visualizer to monitor the robot movements.
- 5.Once a trial inspection is complete, the map, route, sensor feeds, and recorded data are analysed for appropriateness and accuracy. Any required amendments are made. The simulation is run again.
Deployment on site
When on-site, the simulation steps are repeated in the real environment with the real robot being mindful of safety and network coverage.
- 1.The real robot is connected to the platform.
- 2.The sensor feeds are tested to ensure the data is being collected by the platform.
- 3.The robot is teleoperated from the platform to generate a SLAM map of the inspection environment.
- 4.Waypoints are positioned on the real SLAM map.
- 5.A trial inspection is carried out while the dashboard of feeds is monitored.
- 6.Any required adjustments are made to waypoints, the map or sensors.
- 7.The robot can now perform an inspection autonomously, scheduled through the platform.
The Results
4 hrs
sim preparation
12 hrs
on site setup
Unlimited
repeat inspections
OLO provided robot control, autonomous navigation, and a data layer in the browser. With no coding a full, repeatable inspection was achieved.
- Repeatable data from every run, stored for analysis.
- People out of hazardous zones.
- Inspection schedules that don't depend on specialist availability.
- A platform for training and education deployed.
DarbyTech now offer a full autonomous inspection solution. Called QuadRaSense, it includes a quadruped and the OLO platform packaged for next-generation industrial inspection.
The OLO Catalyst programme is designed to help you deploy your first robot on industrial inspection.
“QuadRaSense aims to provide an affordable, interactive education kit for schools, colleges and universities, actively demonstrating how the integration of robotics, AI, wireless IoT sensors and other emerging digital technologies can solve real-world industrial challenges. We have been working closely with OLO Robotics for nearly two years. Their intuitive software platform provides the perfect foundation to start developing these applications—turning a standard remote-controlled quadruped into a fully autonomous, AI-driven inspection and data-gathering system.”
For the of robotics.
Test in simulation. Visualise, teleoperate and script in the browser. No installation required.