ESP32 vs. Fuel Thieves: Digitizing Legacy Trucks in Bangladesh

A Real Problem: Fuel Theft

At Ruhul Engineering, we love building things that solve real problems. And if you talk to any fleet manager or truck owner in Bangladesh, one problem comes up constantly: fuel theft.

For a small business owner, fuel is the biggest operating cost. But in developing economies like ours, most trucks on the road are legacy vehicles—older models with analog dials and no digital brains. Drivers (or third parties) often siphon fuel, and owners have no reliable way to prove it.

Big commercial solutions exist, but they’re built for modern western trucks and often cost $800+ to install plus monthly fees—simply unrealistic for local operators.

So we asked: Can we build something that rivals big systems, but costs less than $50?

The answer is yes. Here’s how we did it.

The Problem: Analog Trucks in a Digital World

The biggest challenge was obvious: we couldn’t just plug into a modern data port. These trucks don’t have OBD-II/CAN-bus. They rely on old-school, float-based resistive fuel senders feeding an analog gauge.

We needed a true retrofit solution—something that non-intrusively intercepts the fuel gauge signal and translates it into clean digital data.

The Solution: The Hardware Stack

We built a custom telemetry unit centered around the ESP32. It’s the brain of the operation. Under the hood:

1) Fuel Digitizer

An auxiliary microcontroller reads the raw analog signal from the fuel sender, filters electrical noise, and delivers clean fuel-level data to the ESP32.

2) Sensor Fusion (GPS + IMU)

We didn’t just want fuel level—we needed context. Using a NEO-6M GPS and an MPU6050 IMU, we can distinguish between fuel drop caused by sloshing on slopes vs. a suspicious drop while parked.

3) Store-and-Forward Architecture

Internet in rural Bangladesh can be spotty. So the device buffers data locally. If the network drops, it keeps recording. When signal returns, it uploads the backlog—so no data is lost.

The retrofit telemetry stack: fuel signal digitization + GPS/IMU context + edge buffering.

Field Test: Does it work?

We didn’t just build this in the lab—we deployed it on commercial TATA LPT 1618 trucks for a 6-week field test.

The results were eye-opening:

• Detection Accuracy: 91.3% recall for theft events
• Cost: ~ $45 total hardware cost (vs. $800+ commercial solutions)
• ROI: Preventing just 25 liters/month saves about $30/month → payback in ~2.3 months

Real-world deployment: installation, calibration, and road testing on legacy trucks.

Why This Matters

This project isn’t just about saving money—it’s about democratization. Commercial tech often ignores the informal sector. But with low-cost components and careful systems engineering, we can bring high-tech fleet visibility to the masses.

Next, we’re exploring Machine Learning (XGBoost) to make theft detection even smarter—filtering out false alarms caused by harsh roads and mechanical noise.

Local problems. Global tech. And a little bit of DIY spirit.