This summer, Adopticum and Magna will collaborate on a development project to explore the utilization of simulation engines in conjunction with game engines to detect and classify unknown objects on the ground. 🤝🔍

The Carla - Adopticum / Magna project aims to explore the integration of simulation engines and game engines to detect and classify unknown objects on the ground. By utilizing the CARLA simulator, we seek to generate simulated and realistic data that can be used to train machine/deep learning models effectively. The project focuses on emulating optical measurement techniques employed by Adopticum and Magna, reducing the dependency on manual data collection.

The major installations are:

• CARLA>=0.9.14 Follow the installation guide for your operating system. The project is developed using the Linux installation. We recommend using the pre-built version of CARLA and installing the additional maps.
• Python>=3.8
• Unreal Engine==4.26 To install Unreal Engine properly you should follow the instructions on the CARLA documentation.

We recommend that you install the Carla environment in a virtual environment and install the Carla Python API via pip. Make sure that you add the root path to your carla installation to CARLA_ROOT. This can be done by adding the following line to your .bashrc file:

export CARLA_ROOT=<path-to-carla>

Make sure to replace the <path-to-carla> with the path to your Carla installation and restart all your terminals.

pip install <path-to-carla>/PythonAPI/carla/dist/carla-<version>-cp<python-version>-cp<python-version>-<os>.whl

pip install ~/carla/PythonAPI/carla/dist/carla-0.9.14-cp38-cp38-linux_x86_64.whl

To install the required Python packages, run the following command in the root directory of the project:

pip install -r requirements.txt

Explain how to use the project, including instructions for running simulations, configuring parameters, and accessing the generated data.

• Simulated and realistic data generation using the CARLA simulator. 🚗🌍
• Training machine/deep learning models to efficiently detect and classify unknown objects. 🤖🦾
• Emulation of optical measurement techniques for accurate representation. 🔫
• Reduction in manual data collection efforts. 📉⏳
• Cost and time savings for data collection processes. 💰

Outline the future plans and potential enhancements for the project. Include upcoming features, improvements, or research areas that will be explored.

• Theodor Jonsson - LinkedIn

• Ruben Forsgren - LinkedIn

• Adopticum 🔬 - LinkedIn
• Magna 🚀 - LinkedIn

The Carla - Adopticum / Magna project is released under the MIT License.

All the major changes can be found in the pages/CHANGELOG.md file. Below you can find examples per described category. Along with some examples of changelogs.

• CARLA 🚗🌍
• Adopticum 🔬
• Magna 🚀