Lab 10

Overview

The purpose of this lab was to understand how Bayes filtering works for localization and implement it in a python simulation.

Code and Explanation

I pulled the stupid move of not using a markdown-based template for my website, which has made my life horrible. To make my life less horrible, I typed most of my writeup for this lab on a Markdown editor and saved it as a PDF, as shown below.

Video

Below is a run of my Bayes Filter! The ground truth is shwon in green, the odometry model is shown in red, and the Bayes filter is shown in blue. The prediction and update statistics are also shown in the video.

The video demonstrates that my Bayes filter worked quite well. A common trend was that error was lower when the robot was closer to a wall, which was likely due to the robot being able to localize better in this case.

References

I referred to Stephan Wagner's website and the course website for this lab. I also discussed ideas with Annabel and Becky.