mgrub/phd_sensor_simulation

Implementation and Documentation of consensus network calibration simulations

Simulation environment to test, compare and evaluate multiple co-calibration methods

Run a given scenario

A scenario can be configured and then multiple methods are applied to the same configuration.

python evaluation_runner.py experiments/scenario_A/

Visualize results of a scenario

Once run, the results can be visualized:

python visualization_runner.py experiments/scenario_A/

Define a scenario

A scenario is defined inside a config.json, which stores information about the

• random state
• reference sensors
• device under test
• measurand
• sensor readings
• methods to be used

All options can also be loaded from an existing (i.e. from a previous run, to achieve or manipulate a specific setting).

Define a method

Available methods are defined in method_args/<method_name>.json.

• class name (as used in the actual source cocalibration_methods.py)
• arguments to initialize the class