At the Massachusetts Institute of Technology, students in the AeroAstro department are confronting one of engineering’s central challenges: designing systems that can make reliable decisions in unpredictable environments. In the course 16.85 Autonomy Capstone (Design and Testing of Autonomous Vehicles), undergraduates work in teams to build, implement, and test complete software architectures for autonomous flight systems. These systems must perceive their surroundings, build maps of unfamiliar terrain, plan safe paths, and land without intervention from human operators. The project simulates conditions comparable to exploring Mars or other planetary surfaces, where a craft must navigate unknown risks and make split-second decisions based on sensor inputs and software logic.
MIT professors Nicholas Roy and Jonathan How developed the course to extend principles from a prior robotics class into the domain of aerial autonomy, emphasizing fault tolerance and confidence in uncertain conditions. Students begin with basic quadrotor drones and no pre-loaded navigation software, forcing them to grapple with the full stack of perception, control, and planning. They then fly these drones through an indoor obstacle course featuring uneven terrain and uncertain landing zones. The hands-on tests are designed to mirror real missions, where unexpected obstacles and environmental uncertainty are the norm.
A central goal of the capstone is teaching students to engineer systems that maintain confidence—the ability to act when predictions are imperfect, and information is incomplete. That involves integrating software with hardware and coping with hidden risks in both the mission and the environment. Team coordination mirrors real-world aerospace projects, with members playing complementary roles to deliver a system that is robust enough to survive unexpected challenges.
By confronting ambiguity head-on, the course prepares future engineers for the reality that autonomy, whether in urban air mobility, reusable launch vehicles, or extraterrestrial exploration, must operate confidently in the face of uncertainty.