Connected and Autonomous Vehicles Simulator

Setting up the environment

To avoid package conflicts it is recommended to create a virtual environment for the dependencies.

Creating the virtual environment

$ virtualenv --python=python3 venv
$ source venv/bin/activate

Installing required python packages

(venv) $ pip install -r requirements.txt

Running the simulation

The simulation can be run through the main.py script in the src/ directory of this project.

(venv) $ python src/main.py

To see the list of all possible parameters to change use --help.

(venv) $ python src/main.py --help

The simulation parameters can also be loaded from a yaml configuration file

(venv) $ python src/main.py --config config.yaml

# config.yaml
simulation:
  length: 150
  lanes: 3
  max-speed: 7
  obstacles:
    - "0:0-10"
    - "2:0-10"
  dispatch: 1

Example usage

Simulation can run in three different modes:

• gui displaying animation of the vehicles
• cli running for a specific number of cycles and showing only statistics
• exp running an experiment for a given list of market penetration rates

GUI

You can see gui specific parameters by running

(venv) $ python src/main.py gui --help

For example to change the animation time of a single step to 200ms

(venv) $ python src/main.py gui --step 200

CLI

You can see cli specific parameters by running

(venv) $ python src/main.py cli --help

For example to change the number of steps to 1000

(venv) $ python src/main.py cli --steps 1000

EXP

You can see exp specific parameters by running

(venv) $ python src/main.py exp --help

For example to set a list of market penetration rates

(venv) $ python src/main.py exp --penetration-list [0.1,0.3,0.5,0.7,0.9]

Changes Report

You can read a report describing the changes in this fork and our experiments at:

https://drive.google.com/file/d/1RKqVRAmWiHqu7a_4HAhg7lrTvW5qTHvN/view?usp=sharing