ASAM OpenMATERIAL® 3D

(The download of ASAM OpenMATERIAL 3D is free of charge)

As the automotive industry relies ever more on virtual methods, the demand for realistic simulations of the real world, or „digital twins,“ is more critical than ever. Accurate sensor simulations require a physically precise representation of the environment, including material properties. ASAM OpenMATERIAL 3D is an open-source initiative that aims to address the need for accurate, standardized 3D assets and material properties to simulate sensor behavior effectively.

The ASAM OpenMATERIAL 3D standard defines physical material properties and standardizes 3D model structures for more accurate simulations. Instead of simply naming materials like glass, metal, or rubber, it precisely defines characteristics such as refraction index, surface roughness, and permeability. These properties are stored in dedicated material files linked to 3D geometries, ensuring realistic behavior in simulations.

Rather than introducing a new 3D format, the standard establishes a consistent hierarchy for model elements like coordinate systems and movable parts, such as a pedestrian’s limbs. This approach enhances compatibility across different simulations and allows existing 3D assets to be seamlessly integrated.

By combining standardized material properties and 3D structures, the standard improves perception sensor simulations, enabling more realistic lidar, radar, and camera outputs. This is essential for generating accurate sensor data, which is critical for testing and training perception algorithms.

The ASAM OpenMATERIAL 3D standard is designed to support key use cases across the entire sensor simulation workflow. It enables the creation of 3D assets and material data, allowing models and material properties to be generated together or separately. These assets and materials can be distributed between organizations through direct exchange or marketplaces, ensuring seamless collaboration.

Once created, they can be imported into simulation platforms, whether they are centralized simulation platforms or decentralized models like functional mock-up units. The standard also supports object part manipulation, allowing 3D models to include movable components within simulations.

By defining material properties, the standard enhances signal propagation modeling, enabling accurate simulation of how sensors interact with the environment, including energy propagation and signal reflection. Ultimately, this facilitates testing and training of perception algorithms, helping refine autonomous vehicle perception systems.

The standard focuses on the virtual environment and objects that sensors will interact with, including

• target objects like vehicles and pedestrians,
• surrounding structures, such as buildings and other fixed elements,
• road surfaces, with physical road characteristics.

However, environmental conditions – such as weather, lighting, and other elements affecting sensors – are outside the scope for now. Similarly, interactions between the sensor housing and the environment, such as raindrops or dirt on the sensor, are not included in this standard.

Standard authors

51Sim Computer Systems Co., Ltd., Ansys Inc., BIPROGY Inc., BMW AG, Continental Automotive Technologies GmbH, dSPACE GmbH, Foretellix Inc., Fraunhofer Institute for Experimental Software Engineering IESE, IAE Suzhou Technologies Co. Ltd., IPG Automotive GmbH, Kanagawa Institute of Technology, Mitsubishi Precision Co. Ltd., Persival GmbH, Pilot Automotive (Shanghai) Co., Ltd., rFpro Ltd., RWTH Aachen, TrianGraphics GmbH, Virtual Vehicle - Das virtuelle Fahrzeug Forschungsgesellschaft mbH, Volvo Group India Pvt. Ltd., WMG University of Warwick