Table of contents {: .text-delta }
Before you Begin
Particle Filters are a direct application of Bayes Filters. The Short Version of Bayes Filter
is shown below:
High Level Overview
We will be doing the following steps:
- Initialize particles randomly across the whole map
- resample at every step (selecting only the particles whose predicted laserscan matches actual laserscan) (low variance sampler)
- Slowly decay the number of particles being randomly reinitialized
- Use the mean of the particles’s estimated pose to get true pose