PIP robot mapping

For my undergraduate capstone, I worked on mapping and navigation for PIP, a mobile robot, in C++, with Python for clustering and visualization.

A small wooden enclosure made a good test because its shape was known, so errors in the map were easy to see. Early sonar readings were noisy, especially in the corners. LiDAR gave a much cleaner outline.

Early sonar measurements shown on a laptop screen — Early sonar capture.

LiDAR point plot outlining the test enclosure — LiDAR outline of the enclosure.

Turning distances into a map

Each reading is relative to the robot. Placing it on a shared map means transforming distance and bearing using the robot’s estimated position and heading, and small errors there make a familiar room look strange.

LiDAR point cloud from a hallway scan — Extending the test to a hallway.

Loop closure and better localization were left as future work. Capstone project, 2022–2023.