Online Adaptive Compensation for Model Uncertainty Using Extreme Learning Machine-based Control Barrier Functions

Requirements

Tested on Ubuntu 18.04

1. Python 3.6
2. Download and uncompress Carla 0.9.10
3. Install package requirements

pip install -h requirements.txt

4. Dowload extra maps and locate it in CARLA_0.9.10/import, then run the following line to install it

   /folder/to/CarlaSimulator/ImportAssets.sh
   

Setup for CARLA simulation

1. Run Carla Simulator

   /folder/to/CarlaSimulator/CARLA_0.9.10/CarlaUE4.sh

   Also set a low quality level if required

   /folder/to/CarlaSimulator/CARLA_0.9.10/CarlaUE4.sh -quality-level=Low

2. Set map from Town06 (the one used in this work) with the following command

   /folder/to/CarlaSimulator/CARLA_0.9.10/PythonAPI/util/config.py  --map Town06" 
   

Usage

There are 5 main files, three used for numerical simulation

src
*
├── main_dummy.py
├── main_elm.py
├── main_nn.py
*

and two for CARLA simulation

src
*
├── main_dummy_carla.py
├── main_elm_carla.py
*

Run any of the codes using the command

python main_*.py