Open-RMF's New Robot Traffic System Saves Warehouse Delays

Open-RMF is like an air traffic control system for robots that share buildings—think delivery robots in warehouses, cleaning robots in offices, or supply robots in hospitals. It prevents collisions by assigning routes and destinations. But the old version had a flaw: if a robot asked for a spot that was already taken, the request just failed and the robot stopped dead. That caused backups and delays.

This new update fixes that. It adds a "reservation" system with parking zones—places where robots can wait when their first choice is occupied. Instead of blocking traffic, a robot is diverted to a nearby charging station or out-of-the-way corner, and its original request stays in a queue. When the spot frees up, it moves in. The result is smoother flow, like a restaurant host seating parties when a table isn't ready yet.

The second fix tackles shared knowledge. In the past, each robot used its own sensors to avoid obstacles, but that information stayed local. If one robot discovered a blocked corridor, others didn't know. Now there's a layered global map: obstacles detected by one robot are shared with the central planner, which replans the whole fleet's routes. It's like a ride-share app learning from every driver that a street is closed and rerouting everyone using the same data.

Building a smarter robot traffic system doesn't just make robots happier—it makes them more affordable and useful. With fewer stalls and collisions, warehouses run faster, hospitals get supplies on time, and offices stay clean without human babysitting. This is open-source work from the Google Summer of Code program, so improvements will roll out to real robot fleets soon.