IoT Telemetry & Fleet Intelligence

FuelShield

Stopping fuel theft before it happens. Real-time telemetry, Kalman-filtered anomaly detection, and sub-second fleet intelligence.

Operational & Financial Impact

20%

Average reduction in fuel expenses (~₹1.2L saved monthly per 10 trucks). Tested across 50+ commercial fleet vehicles, eliminating unaccounted fuel siphoning.

99.4%
Ultrasonic Level Accuracy
< 800ms
Theft Alert Latency
ESP32
Ultrasonic Transducer
NEO-6M GPS
SIM800L GSM
HiveMQ MQTT
Firebase Realtime
Next.js 16
Recharts

The Problem

Fuel theft is one of the most silent yet expensive losses in logistics fleets. Traditional fuel logs and manual dipstick audits fail to detect real-time losses during overnight parking, unauthorized fuel transfers, or partial siphoning.

The Solution

FuelShield was built to make fuel behavior visible in real time. It monitors liquid height via non-invasive ultrasonic sensors, runs edge noise-filtering algorithms, and alerts fleet managers instantly when an anomaly occurs.

Signal Pipeline

How FuelShield Detects Theft

From raw ultrasonic waves inside the diesel tank to instant push notifications on fleet manager devices.

01 // STEP

Ultrasonic Wave Polling

Sensors mounted at the bottom of fuel tanks send high-frequency sound pulses (100Hz) through the diesel liquid layer.

02 // STEP

Kalman Filter & Slosh Reduction

Edge ESP32 runs a Kalman Filter to eliminate noise caused by vehicle vibration, liquid sloshing, and incline variations.

03 // STEP

MQTT Telemetry & GSM Failover

Fuel height & GPS coordinates are packed into binary MQTT payloads and sent via 4G/GSM to HiveMQ brokers.

04 // STEP

Instant Push & Audit Log

If the fuel volume drop rate exceeds threshold outside geofenced pumps, emergency alerts are pushed to Next.js dashboard.

System Capabilities

Non-Invasive Ultrasonic Sensing

Measures fuel height from outside the bottom of the tank without drilling or tank modifications.

GPS & Geofencing Verification

Cross-checks sudden volume increases against known petrol pump coordinates to auto-verify refuels.

Instant Theft Severity Classification

Differentiates between normal engine consumption rates and critical siphoning drops (>0.5L/min).

Fleet-wide Analytics

Interactive charts showing fuel consumption per km, driver efficiency, and historical anomaly logs.

Tamper & Cut Detection

Internal backup battery triggers an immediate alert if the main vehicle battery leads are severed.

Enterprise Security & Sync

Encrypted MQTT over TLS with Firebase Authentication and sub-second multi-device synchronization.

Data from the Tank.

FuelShield bridges embedded hardware and modern web systems. An ESP32-based edge device polls ultrasonic sensors at 100Hz, running a continuous Kalman filter to isolate true liquid height from sloshing.

When abnormal drops are detected, the device publishes an MQTT payload containing fuel volume, timestamp, and NEO-6M GPS coordinates. The Next.js dashboard receives this stream in sub-second time.

Hardware Layer

ESP32 (240MHz) + Ultrasonic Transducer + SIM800L GSM Failover + NEO-6M GPS.

Communication Protocol

MQTT over TLS (HiveMQ Broker) with GPRS & SMS emergency fallback.

Application Stack

Next.js 16 Dashboard + Firebase Realtime DB + Recharts Telemetry Charts.

telemetry_processor.ts
// 1. Initialize MQTT Connection
const client = mqtt.connect(BROKER_URL, {
  clean: false,
  clientId: 'fuelshield_fleet_node'
});

// 2. Telemetry Stream Listener
client.on('message', (topic, message) => {
  const telemetry = JSON.parse(message.toString());
  const { fuelLiters, dropRate, gps, isMoving } = telemetry;

  // 3. Theft Anomaly Detection Engine
  if (dropRate > 0.5 && !isMoving) {
    // Check if truck is at authorized petrol pump
    const isAtGasStation = verifyGeofence(gps);
    if (!isAtGasStation) {
      triggerEmergencyAlert({
        truckId: telemetry.truckId,
        stolenVolume: dropRate,
        coords: gps,
        severity: 'CRITICAL_SIPHON_ALERT'
      });
    }
  }
});