We all have grand visions for our robots, be it straightening our
home, taking out the trash, or even delivering an ice cold beverage.
One challenging prerequisite to these tasks is determining where the
robot is in its environment. If the environment is known then this
problem is called localization. If the environment is unknown then
this problem is called simultaneous localization and mapping (SLAM).
In this talk I will describe the full SLAM problem, I will present an
easy to implement SLAM algorithm, and I will cover some traps and
tricks from my experience implementing SLAM algorithms. I will also
make a brief argument for probabilistically sound implementations. In
the remaining time, and as long as you want to stick around, I will
answer any questions you may have about SLAM in general or about your
own specific implementations.
