A primary objective of this course is to introduce participants to
theoretical and practical issues involved in designing autonomy
algorithms in fielded unmanned marine vehicles. A key insight we hope
to convey in this course is the relationship between effective
autonomous decision-making, sensor processing, and
communications. Often the strengths in one aspect may compensate for
the shortcomings in others, in effect creating an interesting design
space of vehicle properties, where certain points in that space are
more appropriate than others for certain design goals and constraints
on vehicle size, cost, reliability and duration.

\medskip

In this class, participants will be introduced to robot architectures at four
different levels - the robotic middleware level for managing software
packages at the process level, a behavior-based architecture for
autonomous decision-making, a field-control or nested autonomy level
for organizing a system of multiple platforms, and finally
the payload-platform interface that allows us to design our solutions
in a vehicle agnostic manner. 

\medskip

The class has a strong lab component with the goal of providing every
participant the end-to-end skills needed to field actual marine robotic
equipment. These skills include an understanding of the robot
architectures, the effects of band-width and range limited
communications on autonomy algorithms, basic C++ programming skills,
cooperative software development and the use of version control
services, and mastery of the command-line.