Exploring Rule-Based Parking with Carla and ROS

ProjectRepo

Carla is a renowned open-source simulator for autonomous driving research, offering a realistic environment for testing and developing autonomous vehicles. In this project, we present an implementation of a rule-based parking motion using Carla and ROS (Robot Operating System). The parking motion follows an open-loop approach, meaning it relies on predefined rules without information retrieval from sensors, potentially leading to crashes.

Installation

To get started with the rule-based parking implementation, follow these installation steps:

1. Navigate to your catkin workspace:

   cd ~/<catkin_ws>/src
   

2. Clone the Carla parking repository:

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

3. Move to the root of the catkin workspace and build the project:

   cd ..
   catkin_make
   

Environment Setup

Before running the rule-based parking, make sure to set up the environment properly:

1. Run Carla using the command:

   ./CarlaUE4.sh
   

   For more detailed instructions, refer to the official Carla documentation.

2. Launch the Carla ROS bridge:

   roslaunch carla_ros_bridge carla_ros_bridge.launch
   

Running the Rule-Based Parking

To execute the rule-based parking scenario:

1. Source the setup file:

   source devel/setup.bash
   

2. Run the parking script:

   rosrun carla_park park.py
   

Example

The implemented routine can be visualized in the provided GIF, showcasing the rule-based parking motion with a time scale of x5.

Specifications

The park.py script implements a basic parking policy for autonomous cars based on geometric information. The script spawns the following entities:

• One ego vehicle at coordinates (x=61.4, y=-7.62, z=0.05).
• Two vehicles inside the parking location adjacent to the ego vehicle at coordinates (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 potential sensor-based parking policies.
• Carla world set to Town 03 carla_maps.
• All vehicles are rotated 180 degrees to align correctly with the street.

Tested Environment

The rule-based parking has been tested successfully in the following environment:

• Ubuntu 18.04.3 LTS
• Unreal Engine 4.22
• ROS Melodic
• Carla 0.9.7

Feel free to explore and integrate this rule-based parking implementation into your autonomous driving research using Carla and ROS.