As operating mining excavators is often an undesirable job (due to location and work conditions) and the operators often require years of training, solutions other than finding skilled workers on the market are needed for matching the increasing demand of excavation works. The goal of my project is to explore the application of modern control and machine learning methods to the area of automated excavation and manipulation of granular materials. As the cost of controllers and sensors has highly reduced, it is feasible to collect a massive amount of data to achieve precise control and acceptable decision-making. In this project, I will demonstrate the manipulation of granular materials by investigating into techniques, including Reinforcement Learning and deterministic algorithm.

“SuperUROP is an opportunity for me to apply what I’ve learned in class to real-world applications. I’ve taken machine learning courses as well as robotic courses and would like to gain some hands-on experience by investigating the manipulation of granular materials. I hope to propose an applicable system at the end of my SuperUROP, providing an alternative for the excavation industry.”