Proof of Concept · Worker Location

Four Radios in One Badge. Because No Single One Covers a Refinery.

A research proof of concept with Petrobras on an onshore site, testing how people move between open yard, enclosed process areas and confined spaces — where each radio technology stops working and the next one has to take over.

Four radios, one badge
The premise

A Refinery Is Not One Environment. It Is Four.

Open yard with clear sky. Process areas dense with steel structure. Enclosed rooms. Confined spaces where a person may be alone. A single radio technology handles one of them well and the others badly — and a worker crosses all four in a single shift.

The usual answer is to pick the technology that fits the hardest case and accept it being wasteful everywhere else. The premise tested here was different: put four radios in the same badge and let the environment decide which one is speaking.

The stack

Each Radio Earns Its Place by What It Does Badly

Choosing a technology is choosing a set of limitations. The four below are in the same device precisely because their limitations do not overlap — where one falls away, another is already covering.

Cellular
Wide area, outdoor, no infrastructure of ours
Carries position and events from anywhere with carrier coverage, with no anchors to install. It is what makes the badge work on day one, before a single node is mounted — and what keeps a lone worker reachable outside the instrumented perimeter.
Where it stops: Costs battery and a monthly line per device. Useless indoors.
LoRaWAN
Long range, low power, own network
Kilometres of range from a handful of gateways, on a battery budget that cellular cannot match. For a large site, one gateway can cover what would need dozens of BLE nodes — at the price of low bandwidth and coarse position.
Where it stops: Tells you the worker is on site. Not which room.
BLE — Anchors and Mesh
Two jobs, one radio: knowing the zone, and getting the data out
Anchors are BLE beacons fixed to a zone — a corridor, a module, a floor of a structure. The badge hears them, compares signal strength across the ones in range, and resolves which zone the person is standing in. The beacon never knows who passed; the badge works out where it is. Each anchor is mapped to a position on the plant drawing, so a zone read becomes a place a supervisor recognises by name.

The same radio then does a second job: the nodes relay through each other instead of each needing a cable and a gateway. In a refinery, running power and network to every position is often the real cost of a project — mesh removes most of it, and a failed node is routed around instead of leaving a hole in coverage.
Where it stops: Zone resolution, not a coordinate. And the anchor map is only as good as the survey behind it — mount them badly and the zones blur into each other.
UWB
Precision, where precision is actually needed
Sub-metre position from anchors with a known geometry. Expensive per square metre, which is exactly why it should not be the whole site — it belongs in the areas where knowing the room is not enough.
Where it stops: Needs anchors, power and a survey. Deploy it selectively.

Worth being explicit about the anchors, because it is the part most often misread: they are not readers. A BLE anchor transmits and nothing more. It does not detect people, does not store passages and has no idea who walked past it.

The badge does the work: it listens, weighs the signals it can hear, and decides which zone it is in. That inversion is what makes the approach affordable at scale — anchors are battery devices you screw to a wall, and adding a zone means adding a beacon and a line on the plant map, not an installation project.

The part that is actually hard

Putting four radios in one enclosure is an engineering problem. Deciding which one to believe, second by second, is a software problem — and it is the one that determines whether the system is usable.

A worker walking from the yard into a process module will, for a few seconds, be visible to cellular, LoRa and mesh at once, each reporting something slightly different. The platform has to resolve that into one position and one zone, without the operator ever seeing three answers.

What was tested

Five Scenarios, Run in the Field

Not a lab bench. People walking a live industrial site, with the badge deciding which radio to trust as the environment changed around them.

Zone resolution from BLE anchors — beacons mounted across corridors, modules and enclosed areas, each mapped to a position on the plant drawing. Entries and exits recorded as people crossed between them, with the handover between radios happening without the operator seeing it.
Fixed, restricted and dynamic geofences — Areas defined in advance, areas that require a valid permission, and areas created on the fly around a temporary activity.
Man-down and duress — Fall detection and a panic button, with the alert carrying position and the person's identity.
Collective emergency — Mustering by area, with the roll call showing who has not reached a safe point — the case where knowing the zone matters more than knowing the coordinate.
Credential validation at the boundary — Certification and medical clearance linked to the person, checked against what the area requires, at the moment of entry.
Being clear about what this was

A Research Proof of Concept, Not a Deployment

This was a research project. The purpose was to establish whether a single wearable could hand over between four radio technologies across the environments of an industrial site — not to run a plant's safety operation. Presenting it otherwise would be dishonest, and it would surface at the first real deployment.

The commercial device showed real limits. Mechanical resilience under demanding use, display legibility in direct sunlight, and alarm volume in noisy environments where hearing protection is worn. All three are stated in the project report, and all three are hardware selection questions rather than platform questions — a different device changes the answer.

Device choice depends on the phase. What suits a construction phase is not what suits steady-state production, and the economics shift with the number of people carried. That is a scoping conversation, not a catalogue decision.

Why this matters beyond one site

Most Vendors Sell You One Radio and Call It a Strategy

And it works, in the part of the site that radio suits. The gaps appear where the environment changes — which in a refinery is every few dozen metres.

What this project established is not that four radios are always the answer. It is that the question is answerable — that handover between technologies can be made invisible to the person wearing the badge and to the operator watching the screen. For a site with genuinely mixed environments, that changes what is worth specifying.

Compare technologies Connected worker RTLS EHS software Emergency roll call Oil & Gas All case studies

Which radio does your hardest area need?

Not your whole site — the one area where knowing the zone is not enough. That single answer usually decides the architecture, and it is a shorter conversation than a full requirements document.