Mando, autonomous driving with Unity Engine
Development of an ADAS front camera simulator
Performance improvement through future application of ML and physics engines


On the 6th, Unity announced that it had signed a business agreement with Mando for the "ADAS Front Camera Simulator Development Project" on April 27th. This agreement marks the first collaboration between a domestic automotive parts company and a game engine developer, with Mando using the Unity engine to create a virtual environment simulator for ADAS front camera training.

ADAS front cameras act as the eyes of autonomous vehicles, detecting objects ahead of the vehicle and controlling its speed and direction. Simulators replicate thousands of road environments and pre-train the ADAS front cameras to enhance their performance.

Pre-training requires graphics that allow the camera to learn and evolve, and the ability to easily and quickly create diverse simulation environments. Mando uses the Unity engine to create simulation templates with numerous scenarios and design new driving scenarios. Furthermore, the editor function allows for real-time modification and editing of templates, allowing for the reconstruction of virtual road environments.

The two companies will develop virtual sensor modules and enhance compatibility with simulators and other simulation tools. Furthermore, we plan to further enhance simulator performance by introducing machine learning and driving physics engines. In addition to technological collaboration, we will also collaborate on various domestic and international marketing initiatives.

“Producing accurate, real-world road scene instances at scale and using them to refine iterative learning is essential to improving the performance of various sensor peripherals to correctly detect objects,” said Dr. Danny Lange, SVP of AI and Machine Learning at Unity. “Unity’s ability to quickly and easily create high-quality synthetic datasets and simulations is effective for developing and training ADAS sensor tools.”