Fuel theft detection uses fuel-level and telematics data to identify losses that do not match normal vehicle consumption. A common rule looks for a sharp, unexplained fall in tank level and checks where the vehicle was located, whether it was moving, and whether the ignition was on or off.

Fuel data may come from the vehicle’s original sensor, CAN bus, or a separately installed probe. The platform compares consecutive readings, applies a minimum drop threshold, and evaluates the event against engine activity, trip history, refuelling records, time, and geofence. A substantial decrease while a vehicle is parked with its ignition off may indicate siphoning, but the alert is evidence of an anomaly rather than proof of theft.

Fuel can appear to fall because of tank shape, road gradient, sensor noise, temperature change, delayed reporting, leakage, calibration problems, or genuine engine use. Effective detection therefore requires sensor calibration, smoothing rules, suitable thresholds, and enough time to confirm that the level does not recover. When an alert occurs, managers should review its timestamp, location, ignition status, nearby stops, driver assignment, fuel-card activity, and available camera footage. Verified incidents should follow a documented investigation and escalation process. Trends across vehicles, depots, routes, and shifts can reveal recurring exposure, while lockable caps, secure parking, access controls, and driver reporting help prevent future losses.