Visual Navigation for Autonomous Vehicles (VNAV)
MIT OpenCourseWare offers this advanced course on vision-based navigation for mobile robots, self-driving cars, and drones. Lessons cover differential geometry and optimization on manifolds, along with 2-view and multi-view geometric vision for real-time motion estimation, calibration, localization, and mapping. Hands-on labs use mini racecar and drone platforms to put the theory into practice, and the course ends with a critical review of recent research and a team project pushing on the state of the art. Materials include lecture content and lab assignments drawn from an MIT graduate course in Aeronautics and Astronautics, aimed at students who already have a solid mathematical background. As with other MIT OCW offerings, all course materials are free to access, with no certificate offered.