Minor Description
I’m looking at creating a ‘Fleet Management’ application that uses simulated GPS and On-Board Diagnostics aka OBD data.
Here are some initial thoughts:
GPS
Create a simulated GPS sensor that reads in saved real GPS data collected from my Garmin (say from a drive from IBM SVL to IBM research). The GPS sensor will read from a .csv file that contains extracted relevant data from a .gmz file.
Potential applications:
1. Send GPS data to say, Kafka. Potentially, there can be a server application to track the fleet of vehicles in real time (ex. vehicles moving on a map).
2. Speeding monitor – filter for > 70mph
3. Geofence exception monitor if the vehicle is outside the Geofence boundaries (simple definition). This may indicate the vehicle is lost, stolen, or making an unauthorized trip.
4. …
OBD
Create a simulated OBD sensor that reads in made-up data.
Potential applications:
1. Send alert if a sensor value is out of range. For example, engine oil temperature overheating.
1a.. Battery voltage low, ….
2. Send alert for Diagnostic Trouble Codes.
3. Show simple embedded analytics (more complex analytics requires domain knowledge). For example:
3a. Hard braking (>7mph) - send alert to driver and send alert to on server. A metric can count the number of such incidents.
3b. Hard acceleration (> ? mph)
3c. Ignition off & motion detected(GPS) -> possible towing
3d. …
Comments?