The Development of Sentinel: Our OBD-II Technology

When we first kicked off our research into automated Road User Charges (RUC) at RUC on Rails, we knew we had to tackle a major elephant in the room: driver privacy. The industry standard was to just slap a GPS tracker on a vehicle and call it a day, but we didn't think that was good enough for New Zealanders. We believed we could build a world-class distance engine that respected the driver by design, not just by policy. That belief led to the creation of Sentinel, our first proprietary OBD-II hardware device.

The core philosophy behind Sentinel was simple: "Privacy by Architecture." We wanted to build a device that was physically incapable of tracking a user’s location. By removing the GPS chip entirely, we ensured that the device could never reconstruct a trip or profile a driver's behavior. Instead, it was engineered to be a "passive interpreter" of vehicle data, reading the distance directly from the source without injecting any messages into the vehicle's brain.

"We saw the way the market was moving, with everyone seemingly okay with constant coordinate logs and route playbacks," says Adam Johnston, CEO. "But we weren't. We wanted to prove that you could get regulatory-grade accuracy without the surveillance side effects. Sentinel was our way of saying that privacy is a technical requirement, not an optional feature."

Under the bonnet, the engineering effort was significant. Our embedded team spent months working with the ESP32-C3 architecture, writing custom firmware in C using the ESP-IDF. Because the device had to operate in real-time, we utilized FreeRTOS to handle deterministic task scheduling. This allowed Sentinel to parse CAN bus signals at 500kbps, identifying manufacturer-specific speed data with incredible speed and precision.

"It wasn't just about plugging a device in and reading a number," says Paul, Lead Hardware Researcher. "We had to develop a dual-mode measurement system. In its baseline mode, it polls standard OBD-II data, but when it detects higher-resolution signals, it switches to an enhanced passive extraction mode. This meant we could hit less than 0.1% cumulative variance during our local fleet testing. It was a massive win to see that level of precision without a single satellite involved."

Briyarne our CTO team led the validation phase, putting the device through its paces across a massive variety of the New Zealand vehicle fleet. The results showed a ~99% compatibility rate, proving that our "edge computation" approach worked just as well in stop-start Auckland traffic as it did on long hauls down the Desert Road. Because all the math happens on the device itself, there was no need for a constant cellular connection or cloud dependency to get the job done.

While RUC on Rails has since evolved to focus more on digital transaction infrastructure and retail distribution systems, Sentinel remains a vital part of our R&D heritage. It stands as a testament to our ability to solve complex, "hardware-level" problems with a focus on precision, privacy, and no-nonsense engineering. It proved that in the world of RUC, you don't need to know where someone is to know exactly how far they've gone.