Robot fleet management

Robot Fleet Management: What It Covers, What It Misses

Published

a fleet manager tells you where every robot is and what it is doing right now. it does not tell you how often the task actually succeeds, or what changed the week it stopped succeeding. both questions matter once you run more than a handful of robots.

what robot fleet management means :

Robot fleet management is the software layer that runs many robots as one system : it assigns tasks, coordinates traffic, tracks the health of each robot, and gives operators one place to see the fleet and step in.

one robot can be run from a laptop. ten robots sharing aisles, chargers and a shift schedule cannot. fleet management is what turns a group of machines into something an operations team can plan around.

what robot fleet management software does :

  • task allocation : orders and jobs become robot tasks, sent to the robot best placed to do them.
  • traffic and shared resources : robots are kept from blocking each other at intersections, chargers, lifts and doors.
  • monitoring and alerts : battery, position, error states and connectivity, per robot and for the whole fleet.
  • remote intervention : an operator can pause, release or take over a robot that is stuck.
  • updates and configuration : software and settings are rolled out to many robots without touching each one.
  • integration : connections to warehouse and manufacturing systems, and to robots from more than one vendor.

mixed fleets are why interoperability work exists. VDA 5050 defines a common interface between mobile robots and a master control system, and Open-RMF is an open-source framework for coordinating fleets from different vendors in one building.

the question fleet management does not answer :

a fleet dashboard is built around the present : which robot is stuck, which one needs a charge, which mission is late. that is the right view for running a shift.

it is the wrong view for deciding whether a behavior is ready. "97% uptime" says the robots were available. it does not say the pick succeeded 97 times out of 100 under the lighting on the night shift, and it does not say whether last week's update made the docking better or worse.

we compare the two workflows side by side, with the vendors' own documentation, in OORB RunFP vs Formant and InOrbit. the short version : keep the fleet platform for operations, and add recorded-run evidence for the release decision.

fleet management vs fleet supervision :

  • management directs the robots : go here, wait there, charge now. it is the fleet's dispatcher.
  • supervision measures the robots : did the task succeed, under which conditions, and what was different when it did not. it is the fleet's independent witness.

OORB works on the second one. RunFP records every run with the code, model, hardware and conditions behind it, and the Mobile Station adds an outside view of the conditions around the robots. we describe the hardware in robot fleet supervisor.

what to measure across a fleet :

the more robots you run, the more a small per-robot difference turns into a pattern nobody can explain. these are the records that make the pattern readable :

  • task success rate per task and per operating condition, with an interval
  • the code, model, hardware and setup behind every run
  • differences between robots that should behave the same
  • what changed between the last good run and the first bad one
  • conditions the fleet meets on site that nobody has tested
  • a comparison against the deployed baseline before every update

the method behind the first one is in measure robot task success rates, and the second is robot run provenance. uneven motor wear, a firmware version that differs on three units, a sensor topic that drops to 0 Hz on one robot : none of these show up in a fleet average.

rolling out a change to the fleet :

a fleet manager can push an update to every robot in minutes. whether it should is a separate decision, and it needs evidence from before the push.

keep a fixed baseline for each task. run the candidate on a few robots, compare its runs against that baseline per condition, and hold back the cells where it got worse. the workflow is in validate a robot policy update before deployment, and the reasoning behind it is in robot policy evaluation. for the whole path from first pilot to full rollout, see robot deployment.

frequently asked questions :

What is robot fleet management ?

Robot fleet management is the software layer that runs many robots as one system. It assigns tasks, coordinates traffic and shared resources such as chargers and doors, tracks the health of each robot, and gives operators one place to see the fleet and step in.

What does robot fleet management software do ?

It dispatches tasks to robots, keeps them from blocking each other, reports battery, position and error states, raises alerts, supports remote intervention, and distributes configuration and software updates. Many platforms also connect to warehouse and manufacturing systems so that orders become robot tasks.

Is OORB a robot fleet management system ?

No. OORB does not dispatch tasks or route robots. OORB RunFP is the reliability layer that works beside a fleet manager: it records every run with the code, model, hardware and conditions behind it, measures reliability per task and condition across the fleet, and ranks what changed when results drop.

Does fleet uptime prove that a robot behavior is reliable ?

No. Uptime says the robots were available. It does not say how often a task succeeded under a given condition, or whether an update made a behavior better or worse. That takes recorded task outcomes, comparable runs and a baseline to compare against.

fleet management keeps the robots moving. supervision tells you whether the work they did can be trusted.

bring run data from one behavior on your fleet and we will show you how reliable it is by condition, where it fails, and what changed when it did.

Robot failing in the field? Get help now