Crane Retrofit with an OEM Sensor Partnership

The asset owner ran a mobile crane fleet with no telemetry at all. No live location, no utilization signal, no maintenance trail beyond paper logs and the foreman's memory. Planning was done by phone.

The sensor OEM was in a different position. They had a strong portfolio for load, tilt and structural metrics, but no edge story: no gateway, no buffering, no documented path into the asset owner's cloud. Good signals with nowhere to go.

The brief was to retrofit the existing fleet without disturbing who owns what. We would deliver the gateway software, the harsh-environment hardware integration, and a clean handoff into the asset owner's fleet platform. The OEM keeps the sensors and their IP, the operator keeps the assets and the data, and the integration belongs to the operator.

Three sensor protocols on one chassis. Standard J1939 off the drivetrain, raw CAN frames from auxiliary instrumentation, and an OEM-proprietary dialect carrying the load and tilt channels. Left untouched, that means three schemas in the cloud per crane. The gateway had to collapse them into one record on board, so the platform never sees the wiring underneath.

The environment is unforgiving. Outdoor mobile cranes see temperature swings, constant vibration, and water ingress. We specified Owasys hardware rated for those conditions and treated the install as engineering work in its own right: antenna placement clear of the boom, power tapped so it survives ignition cycling, cabling routed and strain-relieved for field service rather than a bench demo.

The partnership boundary was a hard line. The sensor OEM's data stays their IP. We built the integration through adapters the OEM documented and signed off, never by sniffing the bus or reverse-engineering a closed stream. That single decision is what kept all three parties at the table.

Each crane carries an Owasys open-Linux gateway sized for the cabin. On it runs an acquisition layer with three adapters (standard J1939, raw CAN, and the OEM-proprietary stream) that maps every channel into a single per-asset record. When a remote yard drops connectivity, store-and-forward buffering holds the records on local storage and replays them in order once the link returns, so a dead zone costs latency, not data.

Records publish over MQTT to a broker the operator owns, and from there forward into the fleet management platform they already ran. The OEM partner reads the sensor channels they care about through a documented adapter exposed on the operator's side. There is no vendor portal anywhere in the path. The operator holds the broker, the topics, and the retention.

OTA updates arrive on a signed channel under the operator's control: image signature checked before anything is written, and a known-good rollback if a unit fails to come back. A field tech swaps a gateway in roughly 15 minutes using a pre-flashed image and a wiring loom that mirrors the OEM service connector, so the work is reconnect-and-go rather than diagnose-on-site.

We certified before we scaled, on three cranes spanning two vintages. That fixed the protocol adapters against real bus traffic, settled the cabin install pattern, and locked the OEM adapter contract before a single extra unit shipped.

From there the home yard came next, with the operator's own service crew running the installs after watching their first ten. By the regional yards, gateways shipped pre-flashed and crews calibrated locally against reference values captured at the home yard. The schedule ran in three bands: certification in weeks 1 to 2, the home-yard wave in weeks 3 to 6, and the regional yards in weeks 7 to 10.

"Where is that crane?" stopped being a phone call and became a glance at a live view. Utilization, location and structural metrics now feed planning, maintenance and dispatch from one dataset instead of three disconnected ones. The OEM partner got their sensor data flowing into a customer pipeline they neither host nor maintain, which turned a missing edge capability into a selling point. Each of the three parties kept exactly what it brought: the operator its assets and data, the OEM its sensors and IP, the integrator the gateway software.

Crane retrofit is one shape of a recurring problem: equipment whose data sits trapped in an OEM island, and an operator who wants that data without ripping out a platform or signing away ownership. The fix is the same each time. Documented adapters instead of reverse engineering, a broker the customer owns instead of a vendor portal, and source code the customer keeps. It carries to any mobile-machine class with mixed-vintage sensors and an OEM willing to cooperate.

Next steps: industrial routers & gateways, condition & process monitoring, and edge software development.