I did research at my university to help advance the field of machine learning. I wrote object-oriented programs using Pygame and OpenAI's Gym toolkit to train Convolutional Neural Networks (CNN's) and implement Reinforcement Learning techniques to teach objects how to navigate virtual environments. Part of my work focused on enabling Unmanned Aerial Vehicles (UAV's) to autonomously balance multiple success and failure criteria to reach their destination as quickly as possible while avoiding risks.

If successful, this would allow UAV's to be trained in virtual environments that emulate real life conditions, such as delivering a product to a customer quickly while avoiding damage to itself, the product, or its surroundings. It would also increase the capabilities of various UAV's working together, as they will be able to consider multiple success and risk factors when autonomously making decisions and mapping their flight paths.