This repo integrates a DQN with autonomous driving in SUMO environment. The problem is to find a set of actions for all vehicles in the simulation such that the cooperative behaviour of the group results in road evacuation for an emergency vehicle.
The work was published in the RTATM2021 conference titled: "Coordinated traffic clearance for emergency vehicles using rein- forcement learning"

Under supervision of Dr. Mohamed El-Shenawy
melshenawy@zewailcity.edu.eg

For more information on the RL-Sumo model, feel free to contact:

• Alaa Hesham: s-alaahesham@zewailcity.edu.eg
• Waleed Awad: s-wmustafa@zewailcity.edu.eg
• Abdulhady Feteiha: s-abdulhady@zewailcity.edu.eg
  

A parallel effort is exerted to construct a ROS (Robot operating system) model along with hardware equipped autonomous vehicles. For more information on that part, please refer to our Robotics heroes:

• Tasneem Omara: s-tsneem.omara@zewailcity.edu.eg
• Mohamed El-Sayed: s-mohammedelsayed@zewailcity.edu.eg
• Nadine Amr: s-nadine.amr@zewailcity.edu.eg

• The sumo version used is a nightly snapshot of past March 19, 2020. Please refer to sumo for Installation and debugging: https://sumo.dlr.de/docs/Downloads.html#nightly_snapshots
  
• to install the required libraries:
  pip install -r ./requirements.txt
  

cd Anarcho1.3
python Anarcho.py --Train

cd Anarcho1.3
python Anarcho.py --Test

• Anarcho1.3: Single RL Agent + One Ambulance Car
• Anarcho1.4: Mutiple Agents (RL/Auto) + One Ambulance Car, with control over lane busyness, rl_percentage ..etc. Demonstration only, no learning.
• Anarcho1.41: Multiple RL Cars, same as 1.4, but should have independent learners. Still not complete.
• Anarscho1.5: Enables running experiments. Introduces the concept of sample(dictionary) and experiment to produce outputs. Still not complete. Cloned from Anarcho1.41.

--------------------------------------------------------------------------