A graduate course in autonomy for marine vehicles. Covers state estimation and sensor fusion, underwater navigation without satellite positioning, path planning under uncertainty, fault detection and recovery behavior, and multi-vehicle coordination.
The organizing theme is the honest representation of uncertainty. A vehicle operating beyond communication range acts on sensors that drift and a map that may be outdated. A system that recognizes its position estimate has degraded and surfaces is functioning correctly; one that remains confident while wrong is the standard mechanism by which equipment is lost.
Students implement and test algorithms on vehicles in the Marine Robotics and Test Basin, with open-water trials aboard R/V Cascadia Strait where the schedule allows.