Repository files navigation Fingerprinting Ultrasonic Sensors With Machine Learning Hey there! This is my research project for the 2024 REU program at Tennessee Tech University.
Collect data from a series of ultrasonic sensors (4) using Arduino
Each ultrasonic sensor is connected to its own INA219 current sensor, which measures electrical quantities
Data to collect: Distance, Voltages, Current, and Power across the ultrasonic sensor
There were two "types" of distances collected: Distance from a static box, and Distance from a moving box
Static: box was placed at different ranges from sensors and data was collected
Moving: box was placed at different ranges, but was pulled closer (100cm) to sensors by a stepper motor
38,000 samples were collected across ~9 hours (aggregate)
Goal is to identify which sensor a given segment of data is from
Data was stored into a txt files in an undesirable format
Developed python script to combine all txt files and reformat data into 1 clean CSV file
Used PyTorch & Scikit-Learn to build Decision Tree & SVM ML models.
About
Fingerprinting Ultrasonic Sensors With Machine Learning
Resources
License
Stars
Watchers
Forks
You can’t perform that action at this time.