Lost vehicles cost logistics companies $12 billion/year (FBI data).

Our Sentinel Security Suite prevented $2.1M in theft for one client alone.

The Challenge

A logistics company with 500+ trucks needed:

• Real-time location (update every 30 seconds)
• Geofencing (alert if vehicle leaves route)
• Historical playback (where was truck X at 2pm yesterday?)
• Tamper detection (disconnected GPS = alert)

The Hardware

GPS Tracker

• Quectel LG69T GPS module
• 4G LTE modem (cellular connectivity)
• Internal battery (continues tracking if external power cut)
• Accelerometer (detect harsh braking, speeding)

Cost: $45/unit (bulk pricing)

Installation

Wired into vehicle's OBD-II port:

• Power from vehicle
• Ignition status (on/off)
• VIN (vehicle identification number)

The Software Architecture

Data Flow

[GPS Tracker] 
  → 4G LTE 
  → AWS IoT Core (MQTT)
  → Lambda (process & store)
  → DynamoDB (current location)
  → RDS (historical data)
  → WebSocket → Dashboard

Real-Time Updates

MQTT message:

{
  "device_id": "TRK-500123",
  "timestamp": 1705785600,
  "lat": 30.2672,
  "lon": -97.7431,
  "speed": 65,
  "heading": 180,
  "ignition": true
}

Lambda processor:

import boto3

def process_location(event):
    data = json.loads(event['body'])
    
    # Store current location (DynamoDB)
    dynamodb.put_item(
        TableName='vehicle_locations',
        Item={'device_id': data['device_id'], ...}
    )
    
    # Check geofence violations
    if is_outside_geofence(data['lat'], data['lon']):
        send_alert(data['device_id'])
    
    # Broadcast to dashboard (WebSocket)
    ws_client.post_to_connection(
        ConnectionId=connection_id,
        Data=json.dumps(data)
    )

Geofencing Algorithm

Point-in-Polygon Check

from shapely.geometry import Point, Polygon

def is_outside_geofence(lat, lon):
    point = Point(lon, lat)
    geofence = Polygon([
        (-97.75, 30.25),
        (-97.70, 30.25),
        (-97.70, 30.30),
        (-97.75, 30.30)
    ])
    return not geofence.contains(point)

Performance: 2ms per check (fast enough for real-time)

Distance-Based Alerts

"Alert if truck is >50 miles from route":

from geopy.distance import geodesic

expected_location = (30.2672, -97.7431)
actual_location = (lat, lon)
distance = geodesic(expected_location, actual_location).miles

if distance > 50:
    send_alert("Vehicle off-route by {distance} miles")

Historical Playback

Query: "Show me where truck #123 was on Jan 15 between 2-4 PM"

SELECT timestamp, lat, lon, speed
FROM location_history
WHERE device_id = 'TRK-500123'
  AND timestamp BETWEEN '2025-01-15 14:00' AND '2025-01-15 16:00'
ORDER BY timestamp ASC;

Map visualization: Plot path on Mapbox with color-coded speed.

Tamper Detection

Power Disconnection

Scenario: Thief unplugs GPS tracker.

Detection:

• Internal battery sends "power_lost" alert
• Continues tracking for 8 hours on battery
• SMS alert to fleet manager

GPS Spoofing

Scenario: Attacker broadcasts fake GPS signals.

Detection:

• Cross-reference cellular tower triangulation
• Flag impossible movements (teleporting 100 miles instantly)
• Compare GPS timestamp vs system clock

The Dashboard

Built with React + Mapbox:

• Live map with all vehicles
• Click vehicle → details (speed, direction, ETA)
• Geofence zones (green = safe, red = violation)
• Historical playback scrubber

Performance:

• 500 vehicles updating every 30 seconds
• Dashboard renders at 60 FPS
• WebSocket connection handles 1,000+ messages/minute

Cost Breakdown

Per-vehicle monthly cost:

• GPS hardware: $45 (one-time)
• 4G data plan: $5/month (2 GB)
• AWS infrastructure: $0.50/month
• Total: $5.50/month

Compare to commercial solutions:

• Verizon Connect: $25-40/month
• Samsara: $30-50/month

Savings: 70-85%

Real-World Results

Logistics Client (500 vehicles):

• $2.1M theft prevented (3 stolen vehicles recovered in 24 hours)
• 15% fuel savings (route optimization based on GPS data)
• 92% on-time delivery (up from 78% before GPS tracking)

Security Measures

1. Encrypted communication (TLS 1.3)
2. Device authentication (X.509 certificates)
3. Role-based access (drivers see own vehicle, managers see all)
4. Audit logs (who accessed what data when)

Lessons Learned

1. Battery backup is essential. Prevents "blind spots" when power cut.
2. Cellular coverage gaps exist. Cache data locally, sync when reconnected.
3. Speeding alerts annoy drivers. Set threshold at 10 mph over, not 1 mph.
4. Geofences need buffers. GPS accuracy ±30 feet, allow slack.

Need fleet tracking for your business?
Explore Sentinel Security Suite