The RML-SENSOR Benchmark is designed to evaluate the performance of RML processors, especially in microcontroller and IoT device contexts. It simulates data from two types of sensors: a temperature and humidity sensor, and another sensor capable of measuring light intensity and air quality. It also generates a third dataset containing information about various events, offering a diverse range of data for performance evaluation. The tool generates datasets exclusively in CSV format. A notable feature of this benchmark is that the scale factor, when multiplied by 10, indicates the expected result size of the datasets.
- Python 3.10 or higher
No additional libraries are required to run the script.
To generate a dataset, follow these steps:
- Open the
rml_sensor_generator.pyfile. - Modify line 7 to include your desired scale factors. The scale factor determines the number of lines in the generated CSV file. For example:
scale = [15, 25, 55]
- Run the script with the following command:
python3 rml_sensor_generator.py
- Place the provided
sensor_mapping.ttlfile in the generated directories and execute the mappings using an RML processor.
In the default configuration, datasets with scales of 15, 25, and 55 are generated.
- Support additional data formats for dataset generation relevant to IoT devices
- Implement JSON dataset generation
- Implement XML dataset generation
- Develop a Docker container for dataset creation