Convention Center Field Test

Discovery Robotics successfully tested the FX250 autonomous floor-cleaning robot in convention center field tests earlier this month. The FX250 was rigorously tested at a 1,500,000-square-foot convention, conference and exhibition building for one and a half days. Testing included 12 hours on the first day and 2 hours the second day. Our floor cleaning robot was designed specifically for large commercial spaces like convention centers, airports, warehouses, and large format retail centers.

Our system cleaned in a large concrete event hall and then transitioned through a set of doors, cleaning into a carpeted hallway space. We did this multiple times with both teach and repeat mode and using Discovery Indoor Navigation, our autonomous cleaning mode.

What is teach and repeat?
Teach and repeat is a mode of operation in which the operator (either an administrator or service user) teaches the FX250 a path which it will then repeat exactly on demand. The teaching phase involves a human operator manually driving the system to teach it a path that will be stored in the FX250. The stored path is then available to be repeated at any time. After the primary teaching, the FX250 repeats the same path by comparing the current observations with previously recorded sensor data, relying on its onboard sensors. It provides functional navigation, but the robot is not capable of performing any activities that it has not expressly been previously taught.

What is Discovery Indoor Navigation?
In our fully autonomous planning and navigation mode, the FX250 uses LiDAR (Light Detection And Ranging) to map a given area in 2 dimensions. LiDAR is the same technology that autonomous cars use for their navigation and bounces light in the same way that sonar bounces sound waves. Creating a map using LiDAR is accurate and economical, and it enables the FX250 to understand where it is in its environment. The FX250 uses that information to provide the safest, most reliable, and most efficient clean.

What functionality did the FX250 demonstrate?
The FX250 transitioned autonomously from task area to task area even through a set of doors, displaying  its full autonomy. It worked effectively on both concrete and carpeted spaces, using the vacuum functionality powered by its proprietary interchangeable tool system.

Furthermore, the FX250 once again proved it could work efficiently in rooms with glass walls, where light coming in from outside can confuse the visual sensor systems in other commercial floor cleaning robots. The FX250 has an advanced onboard ultrasonics sensor suite that works seamlessly with the LiDAR system to handle any glass areas.

Finally, the convention center space we were cleaning provided rooms that were irregularly shaped and contained unique dimensions and angles. Unlike some other robots that can only navigate within a square or rectangular space, the FX250 proved its capability to confidently clean areas of any shape.

What challenges did the FX250 face?

While the FX250 can capably handle both angles and glass environments, we learned that glass walls with angles from floor to ceiling can inhibit full access to corners and wall edges. Based on our experience at the convention center, we are currently working on an optimization that will allow the FX250 to maximize functional cleaning area in these unique situations. While smaller than many commercial floor cleaning machines, the FX250 does weigh in at about 900 pounds (much of it in our lithium-ion battery system (link) that powers our industry leading 8-hour run times).