Suplementary material for manuscript submitted to Applied Artificial Intelligence:

Autonomous behavior selection for self-driving cars using probabilistic logic factored Markov decision processes

This repository contains the required software to reproduce the results presented in the manuscript paper "Autonomous behavior selection for self-driving cars using probabilistic logic factored Markov decision processes". Such results can be summarized as follows:

• Experiment 1: Test in an environment with 5 static cars
• Experiment 2: Test in an environment with 10 static cars
• Experiment 3: Test in an environment with 5 moving cars

Methodology details are described in the contributed paper.

Requirements:

• Ubuntu 20.04
• ROS Noetic (http://wiki.ros.org/noetic/Installation/Ubuntu)
• Webots 2022a (https://github.com/cyberbotics/webots/releases/download/R2022a/webots_2022a_amd64.deb)

Installation

Note: It is assumed that Ubuntu OS, ROS platform and Webots simulator are already installed.

• $ cd
• $ git clone https://github.com/mnegretev/Autonomous_Behavior_Selection_2023
• $ cd Autonomous_Behavior_Selection_2023/catkin_ws
• $ catkin_make
• $ echo "source ~/Autonomous_Behavior_Selection_2023/catkin_ws/devel/setup.bash" >> ~/.bashrc
• $ source ~/.bashrc

Testing

To test each experiment:

• $ roslaunch aai_2023 experiment1_5_cars.launch
• $ roslaunch aai_2023 experiment1_10_cars.launch
• $ roslaunch aai_2023 experiment2.launch

In each experiment you should see a simulator like this:

Video

BehaviorSelectionExample.mp4

Contact

Héctor Avilés
havilesa@upv.edu.mx
Marco Negrete
marco.negrete@ingenieria.unam.edu