FleetMeshFleet orchestration
The decisions behind a fleet.
A message interface carries information. Orchestration decides how shared work and shared space should be used.
Start with responsibilities
VDA 5050 defines a fleet-control interface, but explicitly leaves traffic-management algorithms outside its scope. MassRobotics describes its interoperability work as information sharing rather than task management. That distinction leaves room for an orchestration policy suited to the actual site.
Sources: VDA 5050 scope and MassRobotics' explanation of information sharing.
A first design review
The following is FleetMesh's engineering checklist, not a prescribed algorithm or a claim about existing FleetMesh software.
| Responsibility | Question to answer | Evidence to retain |
|---|---|---|
| Task allocation | Which available robot can perform the job, with the required capabilities? | Eligibility rules and reasons for rejected assignments. |
| Traffic | Who grants access to a constrained segment, and how is a waiting cycle resolved? | Reservation history and a reproducible contention test. |
| Charging | When should a robot leave service, and which charger can it use? | Energy thresholds, charger constraints and interrupted-job policy. |
| State | How old can an observation be before it is no longer trusted for a decision? | Timestamps, freshness rules and reconciliation behavior. |
| Fault recovery | Who decides whether work is retried, reassigned or escalated? | Ownership, retry bounds and a record of human intervention. |
Work through one shared aisle
Imagine two robots approaching an aisle that your site policy treats as a single shared resource. Write down how the resource is reserved, what each waiting robot sees, and what happens if the current holder loses its connection. Do not begin with throughput claims: first make the state transitions understandable.
Then test the same scenario with delayed state, rejected actions and a robot that becomes unavailable. This is a hypothetical design exercise, not evidence of a FleetMesh deployment.
Keep the protocol boundary visible
A controller can use the protocol's order and state exchanges as inputs and outputs. The choice of scheduling or recovery strategy belongs in a separate design record so that a protocol upgrade does not silently redefine site policy.
FleetMesh recommendation, informed by VDA 5050's division of controller and robot functions.