Virtual reality and 3D scans: Digitally planned production of the Audi e-tron GT

The Audi e-tron GT is the first four rings model for which production was planned entirely without physical prototypes. Multiple technical innovations made this possible, including three-dimensional building scans, machine learning processes and the use of virtual reality. All assembly processes, such as procedures and employee actions were tested and optimized in virtual spaces that model their real-world counterparts down to the finest detail. Virtual planning is now used across site boundaries, enabling digital, connected working without business trips or foreign assignments – and not just during the coronavirus pandemic. 3D scans and the planning in virtual spaces make processes more efficient and sustainable.

Where is virtual planning used and why are 3D scans so important for this?

A variety of prototypes are used in the conventional planning process for the production of a new Audi model. The vehicle prototypes are fabricated during the early planning phase as one-off models with hand-built parts. This is time consuming and cost intensive. Assembly Planning uses these prototypes to define and optimize the later production processes. What are the employee’s tasks? Where does a part have to be located for the employee to have optimal access to it? Can the employee hold and install the part by him- or herself? How does she have to move to do it? Are other parts in the way? What tools does he need? During production planning for the Audi e-tron GT, these questions were derived and answered entirely in the virtual world. Every step and every action was tested in the digital space using virtual reality. The goal of virtual planning is to ensure that during the later production of the vehicle, all processes are perfectly meshed and the cycles along the line are seamlessly coordinated. This requires that every detail of the production hall be modeled precisely and to scale. This is where 3D scans come into play. Using special hardware and software, they create a virtual reproduction of the physical production facility, including all equipment, tools and shelves.

The Audi Böllinger Höfe at the Neckarsulm site, where the Audi e-tron GT is built, thus also exist in the digital world. And thanks to new, digital planning methods, the production of tomorrow can be planned virtually years in advance on the basis of this model.

How do 3D scans work and what role does artificial intelligence play?

A scanner – the hardware – is essential for generating the corresponding data.

It is roughly two meters (6.6 ft) tall and is mounted on four wheels so that an employee can move it around the spaces. At the top is a LiDAR (Light Detection and Ranging) unit and three additional laser scanners as well as a camera. Two processes are conducted simultaneously while scanning a space: The wide-angle camera takes a picture of the space while the lasers precisely measure it and generate a three-dimensional point cloud of the surroundings. 250,000 square meters (2,690,977.6 sq ft) of production hall space at the Neckarsulm site alone has already been scanned using this technology. But it is the interaction between the hardware and software that takes the points, images and data sets generated and turns them into a usable overall image that can be used with the existing planning systems. The software used here is an in-house Audi development based on artificial intelligence and machine learning. The point cloud and the photographs are combined to produce a three-dimensional, photo-realistic space similar to what is seen in Google Street View. Proportions and sizes are true to scale and correspond to reality. The software also automatically recognizes all objects, such as machines, shelves and systems in the space.