openpilot0.8.9

This is an experiment platform with autonomous agent openpilot0.8.9 and simulator carla 9.11 with or without docker setups.

Recommended system

Ubuntu 20.04
NVIDIA RTX 2080, RTX 3070, RTX 3080, RTX 3090 or higher
Python 3.8.10
pip >= 20.0.2

steps to setup the environment

Check the recommended Nvidia driver and install it.

sudo apt-get update
sudo apt-get upgrade -y
ubuntu-drivers devices

Insatll Docker if you want to run the simulation in docker (optional).

sudo apt install curl

curl https://get.docker.com | sh \
&& sudo systemctl start docker \
&& sudo systemctl enable docker

steps to setup the openpilot environment

clone the code from this repository

git clone https://github.com/gitguige/openpilot0.8.9.git
cd openpilot0.8.9

Note: if your files under openpilot0.8.9/models are very small (<5MB), you may need to manually download them again from the repository or install Git LFS to clone te repo again.

extract phonelibs.zip to replace ./phonelibs/
make a virtual environment and activate it

sudo apt-get install virtualenv
virtualenv --python=/usr/bin/python3.8 ~/venvOPnew
source ~/venvOPnew/bin/activate

install dependencies to run the simulation without docker using the autoscript (skip this step with docker setups)

sudo chmod +x setup_autoscript.sh
./setup_autoscript.sh

A manual instruction to install the dependencies is also included in README_install_op-denpendency.md

scons -c && scons -j$(nproc)

install CARLA

Insatall CARLA simulator following the instructions here
install carla lib for docker setups (skip this step if you don't use docker)

source ~/venvOPnew/bin/activate
cd CARLA_0.9.11
easy_install PythonAPI/carla/dist/carla-0.9.11-py3.8-linux-x86_64.egg || true

If the file carla-0.9.11-py3.8-linux-x86_64.egg does not exist, copy one from carla-0.9.11-py3.x-linux-x86_64.egg and rename it to carla-0.9.11-py3.8-linux-x86_64.egg.

steps to run the simulation with CARLA

run carla 9.11 simulator with docker

cd tools/sim/ 
sudo ./start_carla.sh #(with docker)

or without docker: open a terminal under cd CARLA_0.9.11 and run:

./CarlaUE4.sh

open a new terminal and run openpilot threads

./launch_openpilot.sh

open a new terminal and run bridge thread

./run_bridge.sh (cruise_speed_lead) (initial_relative_distance) (new cruise_speed_lead)

steps to run simualations with fault injection

replace the last step above with the following instruction:

./run_fault_inject_XXX.sh (e.g., run_fault_inject_monitor_V2_campaign.sh for random FI)

Context-Aware: use STPA FI campaign 1-6;

Random-ST: use random FI campaign 1-6;

Random-DUR: use STPA FI canpaign 11-16;

Radnom-ST-DUR: use random FI campaign 21-26;