The solar panel voltmeter was created using knowledge of Arduino code and circuitry, transmittance of data achieved through an Arduino Bluetooth Low Energy library which sent the data to the LightBlue app, which forwarded the data to the final display, Adafruit IO. Used Kicad to create sketches of project components ** EDIT LATER A knowledge of troubleshooting was the greatest takeaway from this project, as many apparently simple errors were not as easily resolved as expected.
Collaborated with undergrad team to create a product for determining specific resonance frequencies for different materials using non-contact excitation, splitresponsibility for the pitch video. Examined current Resonance Ultrasound Spectroscopy methods with the help of the Los Alamos National Laboratories(LANL) faculty and postdoc researchers
Worked through tutorials on linear regression, Support Vector Machine, and K-Nearest Neighbors algorithms. Configured a Tensorflow convolutional neuralnetwork using Keras to classify spider, tick, and mosquito bites (>80% accuracy) -TrenchantDroneOps/InsectAI (github.com)
Worked on beautifulsoup module in python for webscraping
The greatest achievement in the Raspberry Pi car that recognized people was triggering the motors in a specific fashion when a person was detected, even though this breakthrough only required a few lines of code. An Arduino was connected to the motor controllers, a Raspberry Pi controlled the Arduino through serial communication, and the Raspberry Pi accepted input from a Picamera module. The Raspberry Pi utilized a basic Tensorflow object detection algorithm able to determine between electronic devices, chairs, people, etc. with more than a dozen categories for classification. The Tensorflow algorithm processed the camera data, and when a person was detected, serial communication was used to run corresponding Arduino C++ code to run the motors.