Application Story

Challenge: Advanced Data Controls Corp has received requests by several Japanese OEM companies to create a photorealistic "camera in the loop" Simulation using a high-performance game engine for ADAS Validation. Thus, a PBR-based (Physically Based Rendering) test system using "Unreal Engine 4" was launched in 2014. As there was no suitable Standard available, we started a research project to implement our Simulation framework with original protocol and various environment models using Unreal Engine.

Two key challenges arose from this research and implementation experience:

1. We needed standardization of interchangeable simulation interfaces for "vehicle dynamics" and "sensor models." As the core of our products, we decided to focus on ASAM-compliant test environments for AV testing solutions.
2. We needed modeling resources to develop high-end PBR-based environment models.

As a solution to 1: We joined ASAM in 2018 and are actively participating in the ASAM OSI project since February 2020. Currently, we are also participating in the standardization of ASAM OpenXOntology, ASAM OpenODD, and ASAM OpenLABEL.

As a solution to 2: We established "VERTechs Inc." in 2016 as an environment model provider for AV virtual testing.

VERTechs is a company that targets the SIL development of ADAS and AD by providing databases and building test environment data to be the best in automotive Simulation. Over four years, we have provided virtual environments for clients in the automotive industry for:

• AI training data creation,
• Reinforcement learning, 
• Virtual sensor models, and Software in the Loop Simulations (SiLS)

More recently, we have added the capability to create geo-referenced data with ASAM OpenDRIVE to include location accuracy to obtain the best possible results that a virtual environment could deliver. With ASAM, we are also looking forward to the standardization of ODD to create environments according to client safety evaluation requirements.

Advanced Data Controls Corp's Vehicle Test Center (VTC/Now on ABD) has 18 years of experience in "Conformance Test Service" for the Japanese automobile industry market in collaboration with c&s GmbH in Germany. In addition, as a simulation domain, we have supported CAN and FlexRay topology simulations with the VHDL-AMS transceiver model and have provided large-scale SIL integration. To create modeling resources to develop high-end PBR-based environment models, we have established "VERTechs Inc." in 2016.

VERTechs has successfully created several virtual environments for OEM companies using game engine technology, taking advantage of high-fidelity graphics and user-friendly GUI. Our virtual environments vary from urban cities to warehouses and vineyards, where autonomous vehicles are necessary for the future. We have also created digital twins of urban cities that range in size from a few hundred meters to as big as 5 kilometers for several OEM companies and Tier 1 suppliers.

VERTechs combined ASAM OpenDRIVE data with a high-fidelity virtual environment for one of our projects: We created a digital twin of the Hamamatsu area located in the south of Japan was created for our customer, an OEM that tested driving scenarios on prescan simulation software. When the OEM entrusted us with this project, the client looked for ground truth to test their driving algorithm. Apart from providing a visually realistic twin city in fbx format to the client, we decided to incorporate ASAM OpenDRIVE data for geo-reference. We used RoadRunner 2019.1.4 for creating OpenDRIVE data and Maya for creating FBX data.

To generate ASAM OpenDRIVE data, we combined information from

• HD maps,
• ADAS maps,
• GIS map,
• point cloud data provided by mapping company.

VERTechs converted ADAS map data into FBX data and ASAM OpenDRIVE data. The total area created was 5 kilometers. 

RoadRunner helped us create a .xodr file previewed with the OpenDRIVE Viewer 1.9.1 (Linux version). With this, ASAM OpenDRIVE helped us provide accuracy in road information through Geo-referencing.

Apart from the road network, we also created other vital elements of the environment necessary for autonomous driving. These included location-specific objects such as foliage, buildings, streetlamps, directional signs, red cones, etc The client then imported the FBX data and the ASAM OpenDRIVE data into Prescan simulation software, and a combination of both solved the localization problem. In the end, the client could successfully drive an autonomous vehicle in our virtual environment with verified data and their choice of simulation software.