Master's Thesis
For my MIT Master’s Thesis I developed a novel method to localize a robotic end effector using the Koopman Operator and image observables. Through my reseach, I designed and built a dexterous gantry crane system where a 6-DOF robotic arm can be inversly mounted to a gantry crane to assist with heavy industry machinery tasks in construction factories. My full thesis can be read here
On the left, we observe a robot with a single camera sensor and a peg, on the right we observe anomoly detection results from a Variational Autoencoder (VAE) model
Gantry Crane with inverted UR-5 robot attached and Realsense D435-i depth camera mounted to end effector