ROBOT LOCALIZATION: PARTICLE FILTER
Implement a particle filter algorithm to perform state estimation and localization to determine the robot's true position on a map
GIF generated from Rvis2 running the robot localizer on a map
In this process I:
• Initialize a particle cloud based on initial 2D pose estimate
• Update particle pose with robot odometry data with Gaussian noise
• Update particle weights with laser scan and based on error of closest mapped object
• Normalize particle weights to sum to one
• Update estimated robot pose with mean particle weights
• Resample particles using a grid style way of taking the mean
After several iterations, the particles converge to a small area to estimate where the robot is.