Case studyRobotics & IoT9 min read

BotYard: one control system for a warehouse running robots from three different makers

A third-party logistics company had introduced autonomous mobile robots in its largest warehouse — from three different manufacturers, each with its own software. Robots blocked each other, tasks were assigned by hand and nobody could see the whole floor. We built BotYard: a unified fleet manager, task orchestration from the warehouse system, traffic management, charging scheduling and live floor visualisation.

19
Features shipped
4
Apps & platforms
6
Team members
22
Weeks to rollout
BotYard fleet control room with live warehouse map
BotYard: every robot, every task, one floor.
01

Three robot fleets that didn't talk

BotYard's head of automation had bought robots from three manufacturers, each best at something different: goods-to-person shelves, tote carriers and pallet movers. Each came with its own fleet software. Robots from different vendors blocked each other in shared aisles, and supervisors assigned work across three screens by hand.

“We bought robots to reduce manual work. We ended up manually managing robots.”

— BotYard's head of automation
02

Watching the floor

We spent shifts on the warehouse floor and in the control room. Congestion happened at predictable points — shared aisles and charging stations. Robots charged at the worst times because each fleet scheduled charging independently. The warehouse management system released work without knowing which robots were available.

What was going wrong
  • Robots from different vendors blocking shared aisles
  • Charging scheduled independently per fleet
  • Work assigned manually across three screens
  • No single view of the floor
Industrial robotic arm
Shifts on the floor revealed where the fleets collided.
03

One brain above three fleets

BotYard sits above the vendor systems, talking to each through adapters and open standards where supported. It receives work from the warehouse management system and assigns tasks to the best available robot from any fleet. A traffic manager reserves shared zones so robots from different vendors never contest the same aisle, and charging is scheduled across the whole floor around demand peaks.

Supervisors see every robot live on one map and handle exceptions — blocked paths, faults, manual interventions — from a tablet.

04

The team

Real-time systems and robotics integration experience led the team.

1
Engagement lead

Floor observation and vendor coordination.

1
Product designer

Control room map and supervisor app.

2
Backend engineers

Orchestration, traffic and charging scheduling.

1
Robotics integration engineer

Vendor adapters and standards.

1
Data engineer

Telemetry and fleet analytics.

6 people in total, working as one team.

05

Decisions we made

Agreed with the head of automation and robot vendors.

01

Replace vendor software or sit above it?

  • Replace vendor fleet software
  • Orchestrate above it through adapters

Our call: Orchestrate above it through adapters. Vendors' low-level navigation worked well; BotYard coordinated across fleets without replacing it.

02

How to avoid congestion?

  • Let each fleet navigate independently
  • Reserve shared zones across fleets

Our call: Reserve shared zones across fleets. Zone reservation prevented cross-vendor blocking.

03

Cloud or edge?

  • Cloud orchestration
  • On-premise edge servers

Our call: On-premise edge servers. Robot coordination needs low latency and must work without internet.

06

Every feature, module by module

Everything that shipped for the control room, supervisors and integration.

Orchestration
One brain for every robot.
  • 01WMS integration

    Work received from the warehouse system.

  • 02Cross-fleet task assignment

    Best available robot from any vendor.

  • 03Priorities and deadlines

    Urgent orders first.

  • 04Vendor adapters

    Each fleet connected through adapters.

  • 05Open standard support

    VDA 5050 where supported.

Traffic and charging
No more blocking.
  • 06Zone reservation

    Shared aisles reserved across fleets.

  • 07Rerouting

    Paths adjusted around congestion.

  • 08Charging scheduling

    Planned around demand peaks.

  • 09Battery monitoring

    Levels across every robot.

Control room
The whole floor at once.
  • 10Live floor map

    Every robot, path and task.

  • 11Exception handling

    Faults and blocked paths flagged.

  • 12Supervisor tablet app

    Handle exceptions on the floor.

  • 13Manual override

    Pause, redirect or remove robots.

Analytics and platform
Measured automation.
  • 14Fleet utilisation

    By vendor and zone.

  • 15Throughput analytics

    Tasks per hour and cycle times.

  • 16Congestion heatmaps

    Where robots wait.

  • 17Telemetry storage

    Time-series data for every robot.

  • 18Role-based access

    Operators, supervisors and engineers.

  • 19Audit trail

    Overrides and interventions recorded.

BotYard fleet analytics
Fleet analytics: utilisation, throughput and congestion by zone.
07

Rollout

BotYard took control of one zone first, with vendor systems still running underneath. Once traffic management and task assignment proved themselves, control extended zone by zone to the full floor.

  1. Weeks 1–3
    Discovery

    Floor shifts and vendor system review.

  2. Weeks 4–6
    Design

    Orchestration, traffic and control room.

  3. Weeks 7–16
    Build

    Adapters, orchestration, traffic and analytics.

  4. Weeks 17–19
    First zone

    BotYard controls one zone.

  5. Weeks 20–22
    Full floor

    Zone-by-zone expansion.

08

What we learned

Orchestration beats individual optimisation. Coordinating fleets did more than improving any one of them.

Adopt standards where they exist. Open interfaces made adding new robots easier.

Built with
  • Next.js
  • Go orchestration services
  • MQTT
  • PostgreSQL
  • TimescaleDB
  • VDA 5050 and vendor APIs
  • Real-time floor map
  • On-premise edge servers
CTA Background

Building something like BotYard?

Bring the problem as it actually is, constraints included. You will get a straight answer on whether we are the right team.

View Our Work
AI-First Engineering
Secure & Scalable
Built to Deliver Impact
More case studies