Custom Audio Machine Learning for Robotic Design

One of my advisors, David Goedicke, does extensive research on how audio can be used to improve human-robotics interactions through machine learning. I spent a large portion of a summer programming with him to create a streamlined system that allows anyone to upload their own audio ML models and use them to control a small robot.

Through node-red, Google's Teachable Machines, and some backend development, we provided a simple process for building robots that naturally react to sounds in ways not possible to the average, non-technical person.

Links to both custom node-red nodes can be found on my GitHub.