Autonomous vehicles have made significant strides in the last few years, moving from their infancy to mature industrial products. Self-driving industries, including Waymo, Cruise, Zoox, Tesla, Argo AI, etc. have been offering various levels of automated driving. In this class, we will dive into the technical stack, understanding the details of each component, including sensing, perception, motion prediction, planning and control, mapping and localization, simulation, and evaluation. Besides this modular approach, the class will also introduce alternative end-to-end approaches, and frontiers of autonomous driving research.
In this course, students will get a chance to learn and apply state-of-the-art self-driving software stack in photorealistic simulators, on real-world autonomous driving datasets, on self-collected data using provided automotive-grade sensors (including LiDARs, cameras), as well as on research vehicles for potential pilot final projects. The class is a combination of theory and practice, which allows students to apply each module they learn from the lecture to an actual experimental prototype. Students will also combine everything taught in the class for a final project, which provides a playground for innovation beyond the state-of-the-art.