Unit 11: Localization, Odometry, and Frames

"Where am I?" is an estimation problem

A robot rarely receives its exact pose from one perfect sensor. It estimates pose from evidence that has timing, uncertainty, and coordinate frames.

Position Is an Estimate introduces pose as data with a source. Odometry and Drift shows how small errors accumulate. Coordinate Frames and Sensor Fusion prevents coordinate values from becoming meaningless numbers.

Draw the frames before debugging them

Keep a pose/odometry trace plus a coordinate-frame diagram.

Label each frame. Mark which transforms are fixed by the robot's construction and which are estimated during operation. Then identify one source of drift and one external observation that could constrain it.

Pose, heading, odometry, drift, transform, frame, uncertainty, and fusion should let you explain the estimate's architecture.

Precision is not certainty

A pose printed to six decimal places is not automatically accurate. Your evidence should preserve where the estimate came from and what can make it wrong.

This Unit can be completed entirely from supplied traces or simulation.