Simulates an electric vehicle's sensor data over time, with synthetic failure modes. See the Avro schema for more info on the data stored in a sample.
Note: Ensure you are using at least Java 8
./generate_test_data.sh
You can also specify the number of data points per car and the number of cars to simulate, e.g.:
./generate_test_data.sh 50 20
for 50
data points and 20
cars to simulate.
The file will be generated as car-sensor-data.csv
in the current directory.
The device simulator can simulate this model with multiple models, publishing their data to an MQTT broker with its behavior configured in scenario.xml
.
You can build the Docker image using build-generator.sh
. This will add the MQTT payload generator to the device simulator base image.
For probabilities of events and implementation details, see CarModel
The car's "shock performance" is measured by accelerometers on the chassis at every corner of the vehicle. In the event of a bump, minor acceleration takes place.
With a certain probability, the shock of a car might fail, in which case, the default acceleration at each iteration increases slightly and the acceleration in case of a bump is much larger.
We assume that the acceleration data here is not a simple point in time measurement, but an aggregation over time.
The car might lose tire pressure with a certain probability. In this case, the pressure will drop below the assumed nominal range of [30..35]
psi.
The car has vibration sensors which monitor the vibration of the engine.
With a certain probability, a car's vibration might increase greatly, indicating that some sort of mechanical degradation or failure is taking place.
The car has a coolant temperature sensor. The fictional (liquid) coolant in this model is used for engine and battery cells.
With a certain probability, a car's coolant temperature might increase greatly (instead of linearly with current draw), indicating battery cell failure , short circuit or other electrical failure.
The rate of decrease of the battery's percentage might increase greatly, indicating battery cell degradation.
Firmware version is indicated in the model, a single car might use an outdated software version, e.g. due to not being online for a while.