Introduction

This assignment involves implementing a ROS package for Monte Carlo localization. The completed code should be (more or less) a drop-in replacement for the default ROS localization package.

Part 1 - Toy Particle Filter

Complete the unfinished methods in particle_demo.py. You should be able to use the completed script to re-calculate approximate answers for the grid-based localization question from HW2.

You may also test the correctness of your methods by running the tests in tests.py.

Submit your completed file through Web-CAT.

Part 2 - Monte-Carlo Localization

I am providing a partial implementation, including a fully-functional Particle class. You should get started by cloning the appropriate github repository.

The package is organized as follows:

particle_filter/
    package.xml
    CMakeLists.txt
    launch/
        all_but_localization.launch <- launches simulator etc.
        everything.launch           <- launches everything (except teleop.)
    worlds/
        bricks.world             <- Gazebo simulation world for testing. 
    rviz/
        mc_localization.rviz     <- Rviz configuration file.
    maps/
        bricks_map.yaml          <- Pre-made map of bricks.world
        bricks_map.pgm
    scripts/
        map_utils.py     
        mcl_tf.py        
        particle.py
        mc_localizer.py          <- UNFINISHED

The four Python files in the scripts folder serve the following roles:

• map_utils.py - This is utility code for working with ROS OccupancyGrid messages. You shouldn't need to modify this code.
• mcl_tf.py - This node broadcasts an appropriate transform from the odom to the map coordinate frame based on the pose estimate provided by mc_localizer.py. This page on ROS coordinate frame conventions describes the relationship between the odom and map frames.
• particle.py - This file contains the definition of a Particle class that may be used in the mcl implementation. You should not need to modify this file, though you are welcome to do so if you want. You should read this file carefully so that you understand how to use the Particle class.
• mc_localizer.py UNFINISHED. This should provide a complete implementation of Monte Carlo localization. Read the Pydoc comments for specific requirements.