In steer-to-pose we implement a differential drive robot (two degrees of freedom) and use a simple control law to drive the robot toward a goal pose:
In piano-mover we compute the configuration space for a robot living in (x, y, theta) space and, rolling our own Dijsktra implementation for planning in voxel space, compute a shortest path from start to goal:
In bg_subtraction we try out a simple mixture-of-gaussian background subtraction algorithm from the opencv library.