vignif / carla-parking

carla simulator for autonomous driving research

Carla parking with ROS

Carla is an open source simulator for autonomous driving research available at http://carla.org/

This package contains an implemantation of a rule based parking motion. It is an open-loop approach for the task, given that no information are retrieved from sensors and crashes can occur.

The example spawns 2 cars in predefined positions (leaving enough space between them) and the ego vehicle is blindly controlled to fit the parking spot.

Example

this gif shows the implemented routine with a time scale x5

Installation

<catkin_ws> = name of your catkin workspace

cd ~/<catkin_ws>/src

git clone https://github.com/vignif/carla_parking.git

cd ..

catkin_make

Environment for running

run carla with ./CarlaUE4.sh as from https://carla.readthedocs.io/en/latest/start_quickstart/

run carla ros bridge with roslaunch carla_ros_bridge carla_ros_bridge.launch

Running

source devel/setup.bash

rosrun carla_park park.py

Specs

park.py implements a basic parking policy for autonomous cars based on geometric information This script spawns:

• one ego vehicle in x=61.4, y=-7.62, z=0.05
• two vehicle inside the parking location a side of the ego vehicle respectively in x=60.4, y=-10.62, z=0.05 and x=47.0, y=-10.62, z=0.05
• a camera attached to the ego vehicle for future sensor based parking policy
• Carla world set to Town 03 carla_maps

all the vehicles are rotated of 180' in order to be correctly in line with the street

tested env:

1. Ubuntu 18.04.3 LTS
2. UnrealEngine 4.22
3. ROS melodic
4. carla 0.9.7

License

MIT license

• Copyright 2021 © Francesco Vigni