A Middle Way to Traffic Enlightenment: Repeatability Evaluation Package
Our repeatability evaluation package includes data, python notebooks for recreating our quantitative figures, and a docker image to playback the recorded ROS bag file simulating the ROS message passing network. If you have never used these before, install the proper resources on your machine here:
- Docker Install: https://docs.docker.com/engine/install/
- Jupyter Install: https://jupyter.org/install
A. File Resources
You can find the relevant files at this Zenodo DOI: https://zenodo.org/records/10611821
B. Docker Image
The Docker image provides a ROS integration to automatically perform a realtime replay the rosbag file from the experiment. To replay the data, navigate to the directory where you downloaded your Dockerfile, and start.sh.
If you are on MacOS or Linux: ensure to: chmod u+x start.sh before running docker build. If you don't do this, you will find that there is an error, /ros_entrypoint.sh: line 6: /start.sh: Permission denied
Then build and run with the following commands:
- $ docker build -t middleway .
- $ docker run -it middleway
The Docker container invokes its own roscore, so external ROS systems may connect to the container as a ROS host to view and analyze ROS topics. Alternatively, The container can connect to an external ROS host by connecting to the container with bash, then manually configure the host and manually perform the rosbag play:
- $ docker build -t middleway .
- $ docker run -it middleway bash
- $ export ROS_MASTER_URI=“http://<external hostname or IP>:11311"
- $ rosbag play 2023_10_23_11_13_42_2T3MWRFVXLW056972cbf_codegen_test.bag
C. Reproducing FiguresInside the Zenodo resource, there are seveal ‘.ipynb’ jupyter notebook files inside directories labelled with the figures they recreate. Figures 6-12 are available to recreate, as the earlier figures in the paper are photographs or diagrams. Once you open the notebooks in Jupyter, you will be able to run a series of small cells with python scripts which will recreate the figures in the paper with the data files inside the folders.