Skip to content

Latest commit

 

History

History
31 lines (25 loc) · 1.59 KB

README.md

File metadata and controls

31 lines (25 loc) · 1.59 KB
Raw

RML-SENSOR Benchmark

Overview

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.

Requirements

  • Python 3.10 or higher

Installation

No additional libraries are required to run the script.

Usage

To generate a dataset, follow these steps:

  1. Open the rml_sensor_generator.py file.
  2. 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]
  3. Run the script with the following command: 
    python3 rml_sensor_generator.py
  4. Place the provided sensor_mapping.ttl file 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.

Planned Improvements:

  • 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