Intro

What is the Neato Runner?

Our project goal was to deploy a Neato runner, optimizing for speed, inspired by the hit game Temple Run. The runner is meant to navigate an unknown environment as fast as possible avoiding collisions. The runner utilizes the concepts of frontier exploration and dynamic path planning.

How to run the Neato Runner?

Clone the repository into a parent directory.
Run ‘colcon build --symlink-install’ in the parent directory to build the project.
Run ‘source ~/ros2_ws/install/setup.bash’ to refresh your terminal’s environment variables.
Run ‘ros2 launch neato2_gazebo neato_gauntlet_world.py’ to launch the Gazebo environment associated with the CompRobo 2024 course.
Run ‘rviz2’ to open rviz2 (optional).
Run ‘ros2 run neato_temple_run depth_estimation.py’ to start the visual and LIDAR scan pipelines.
Run ‘ros2 run neato_temple_run main.py’ to run the Neato Runner.

How to unbag the evidence?

In the parent directory, run ‘ros2 bag play src/neato_temple_run/bags/gazebo_full_system_test_2/ --clock’ in order to play the bag of all system elements. The ‘gazebo_full_system_test’ bag contains no saved publishers other than the Neato’s.
Run ‘rviz2’ to open rviz2.

What dependencies do I need?

CompRobo 2024 Class Setup: https://comprobo24.github.io/How%20to/setup_your_environment
For RRT:

A comprehensive list of packages included in the work environment are listed in the ‘requirements.txt’ file. There are likely some unused packages listed, but most are used in some way or another. The most important packages to pay attention to are NumPy, SciPy, MatPlotLib, and ContourPy because these must not be upgraded to version 2 and beyond to retain compatibility with the ROS Humble environment.

For Vision:

Make sure that PIL is installed.
Install Transformers ‘pip install transformers’.
Install PyTorch.
Follow this to install pytorch. We found that the CPU only installation gave sufficient results.