AMR Controllers: The Brownfield Fix UK Engineering Ops Miss

AMR controllers are the onboard control units that turn a mobile robot's sensor feed into motion, safety decisions and fleet-level instructions — the layer that sits beneath every autonomous forklift and AMR on a plant floor. The stakes are physical: of the 126 workers killed in work-related accidents in Great Britain during 2025/26, 24 were struck by a moving vehicle and 21 by a moving object, according to HSE's provisional figures. If you run operations for a UK engineering or heavy-parts site, you already feel the gap. You have a brownfield plant, irregular loads — castings, engine sub-assemblies, transmission cases — vehicles of four different ages sharing one aisle, and a successful robot pilot in one bay that will not scale into the next. The reason is rarely the robot. An AMR controller is the single component that sets a mobile robot's safety rating, its navigation method, and whether a fleet manager can command it alongside vehicles from other manufacturers.

Why engineering sites stall after the first successful robot

Heavy-parts plants, from the West Midlands to the industrial estates around Burton-on-Trent, are the hardest brownfield automation case in UK industry, for structural rather than technical reasons. Aisles laid out for people narrow unpredictably around press lines and work-in-progress buffers. Loads are machined castings, transmission cases and part-built sub-assemblies with shifting centres of gravity, not uniform pallets. And the fleet was bought over fifteen years, in separate capex rounds, from whichever channel was cheapest that year.

That history defeats the second phase of automation. The first robot works because it is commissioned as an island: one route, one load type, one set of hand-drawn safety boundaries. Add a second vehicle from a different source and you discover that neither unit will accept instructions from the other's control software, that their obstacle-detection behaviour differs at the same closing speed, and that no single screen shows you where both machines are.

Regulation sharpens the problem. Workplace transport remains one of the most persistent causes of serious harm in UK industry — HSE recorded 59,219 non-fatal employee injuries under RIDDOR in 2024/25, with being struck by a moving object accounting for around a tenth of them. HSE's workplace transport guidance expects segregation, visibility and controlled interaction between vehicles and people. A fleet you cannot see as one system is a fleet you cannot evidence as safe — and that is what stops an operations director signing off bay two.

An AMR controller is the single component that sets a mobile robot's safety rating, its navigation method, and whether a fleet manager can command it alongside vehicles from other manufacturers.

Lever one — specify the control layer before you specify the fleet

Reverse the usual procurement order. Most sites choose vehicles and inherit whatever control units arrive inside them, which is how a plant ends up with four incompatible stacks. Decide the control standard first and make vehicle selection answer to it.

Three requirements go into the specification before any machine is quoted: the controller must expose an open fleet interface, carry a stated safety performance level, and support the navigation method your building actually allows. A press shop with reflective surfaces and constantly moving work-in-progress needs natural-feature navigation, not a fixed guidance path someone will block within a week.

This collapses a surprising amount of lifetime cost. One controller family means one spares holding, one engineering competency, one set of commissioning tools and one integration to your business systems rather than a fresh project each time. It also preserves your freedom to buy the best vehicle for each task later. Our AMR controllers range is built on that principle: a common control standard that different vehicle classes can be built or retrofitted onto.

Lever two — make one fleet manager the arbiter of every movement

This is the technical lever, and it turns a pilot into an operation. Vehicle-level control decides how a single machine moves; it cannot decide which of six machines takes the next job, who yields at the junction outside the paint line, or how charging is sequenced. Those are fleet decisions, and they need a layer above the vehicles.

What makes that possible across a mixed-age, mixed-source fleet is VDA 5050, the open standard for communication between mobile robot fleets and a central fleet control. It uses MQTT transport and JSON messages, so one fleet manager can issue orders to vehicles from different manufacturers in a common format instead of needing a bespoke interface per supplier. Where a controller speaks it natively, integration becomes configuration rather than development.

Work should also come from the systems you already run, not a parallel schedule maintained by hand. M4 fleet manager converts tasks into routes, traffic rules and vehicle assignments; RDS robot dispatch handles allocation and priority. Both take work from your existing production and warehouse systems through a documented interface, as set out in our approach to ERP and WMS integration, and write completions back so those systems stay the record of stock.

Lever three — build the PUWER case into the controller choice

The regulatory lever, and the one most often deferred until it is expensive. Under the Provision and Use of Work Equipment Regulations 1998, duties for safe work equipment sit with the employer operating it. HSE's PUWER guidance requires equipment to be suitable for its use, properly maintained, and fitted with adequate controls and protective devices. An autonomous vehicle transfers none of that to the integrator.

In control terms, what satisfies it is a safety-rated architecture: a controller with a declared functional-safety performance level, safety laser scanners whose protective fields change with speed and load, and emergency stop circuits inside the certified chain rather than bolted on. ISO 3691-4, the international standard for driverless industrial trucks and their systems, is the reference assessors expect you to have worked to, alongside the ACOP guidance in HSE's L117 for rider-operated lift trucks where manual and autonomous vehicles share space. UKCA marking and the site's own risk assessment complete the file.

The practical consequence is that safety-rated control cannot be retrofitted cheaply. A functional-safety forklift controller specified at the outset costs a fraction of re-engineering a fleet commissioned without one, and it is what lets autonomous and manned trucks share an aisle instead of a fenced cell.

Lever four — retrofit the assets you already own

The commercial lever. Engineering sites typically hold serviceable trucks with years of life left, and replacing them wholesale to automate is rarely the best use of capital. Brownfield retrofit — fitting an autonomous forklift controller, navigation sensors and safety scanners to a suitable existing vehicle — converts an asset you have already paid for.

Retrofit is not universal. The base vehicle needs electric drive, an accessible control interface and sound mechanical order; a truck near end of life should be replaced, not converted. Where it applies it shortens lead time, because the machine is already on site and known to your maintenance team.

For the mixed case — some retrofits, some new FlyWei autonomous forklifts, some lifting robots for sub-assembly moves — phasing the cost matters more than the headline price. FlyWei leasing over 3, 5 and 7-year terms lets a plant roll out bay by bay against operating budget instead of waiting for one large capex round — usually what keeps a stalled pilot stalled.

What FlyWei does here

FlyWei is an independent, vendor-neutral UK systems integrator of autonomous forklifts and AMRs. We integrate the best machines across multiple manufacturers rather than selling one range, which matters in a brownfield plant because the right answer is almost always a mixture: a retrofitted counterbalanced truck on the heavy-casting move, a stacker on mid-height racking, a heavy-lift AMR shuttling engine sub-assemblies to the line, and manual trucks where they earn their keep.

The control layer is where we start. FlyWei engineers survey the site, confirm which existing vehicles are credible retrofit candidates and which movements justify a new machine, then set one control standard across the fleet — safety-rated where vehicles and people share space, VDA 5050-capable so M4 can command every unit regardless of origin.

From there M4 handles traffic, routing and charging strategy across the fleet, and RDS allocates jobs from your existing production systems. Commissioning is phased bay by bay, keeping production moving. Support comes from UK-based engineers, with the control architecture documented well enough to stand up in a PUWER review. Our earlier piece on AGV controllers in co-packing operations covers the same architecture in a contract-packing setting, and Logistics UK publishes sector guidance on fleet operations.

Frequently asked questions