Robot Autonomous Racing DeCal is a course dedicated to teaching students to gain a holistic view and hands-on experiences about building a high-performance fully autonomous vehicle from scratch. In this semester, due to physical limitations, the focus shifts to learning the algorithms behind autonomous racing/driving with the help of our simulation platform (so now students do not need to construct the cars themselves). The one-semester course is built on the ROAR platform and its racing events in the EECS Department. Students will learn

● The history of the development of modern autonomous driving technologies.
● Global and local planning algorithms based on occupancy grid maps.
● Real-time 3D computer vision algorithms to parse road conditions and autonomous control algorithms to race on a track and avoid obstacles.
● Control algorithms that enable the vehicle to perform complicated tasks such as waypoint-following and lane-following.
● And other interesting topics surrounding robotics, planning -- e.g. reinforcement learning and neural networks.

At the end of the semester, students will team up using the algorithms they learned and developed to participate in the virtual ROAR competition, hosted by the EECS department.

Mode of Instruction: Pre-recorded Lectures with Synchronous OH
Structure of Grade (tentative): 25% Checkpoints, 50% Homework, 25% Final Competition (Participation)

[UPDATE 8/25: application is closed for now due to the high volume of applications. If you, for some reason, really want to take this course, pls send an email for more info] To enroll in the course, you first need to fill out the interest form attached below. In the first week of school, we will release an info session video in which we will specify an action item that you need to do in order to enroll (this action item is not meant to challenge you but to make sure you meet some hardware requirements for this course). We will send permission codes to students who satisfy the requirements.