Virtual Homologation of Driver Assistance Systems: A Škoda Innovation

Virtual Homologation of Driver Assistance Systems: A Škoda Innovation

Computer simulations are now routinely used by carmakers in vehicle development. However, Škoda Auto has developed a unique solution in the field of driver assistance systems that also helps with vehicle approval for road use, known as homologation. The brand has already applied for a patent in connection with this innovation.

16. 11. 2026 Škoda World

When testing the now mandatory Intelligent Speed Assistance (ISA) function as part of the vehicle homologation process, Škoda Auto uses a methodology known as virtually assisted homologation. This allows the company to carry out most of the tests on computers. The results of virtual tests are just as conclusive as those of real-world tests. As a result, the carmaker saves dozens or even hundreds of weeks of demanding real-world testing every year. 

A First Step 

When developing the system, specialists from Škoda Auto and its subsidiary Digiteq Automotive drew on their previous experience with the company's internal virtual testing of camera systems. Škoda uses this approach to test, for example, its High Beam Assist, traffic sign recognition, lane keeping and other ADAS functions (Advanced Driver Assistance Systems). Intelligent Speed Assistance proved to be an ideal candidate for virtualising a significant part of the homologation process. Recent developments in European legislation have also contributed to this, as they increasingly allow physical tests to be replaced by virtual ones.

Navrh-bez-nazvu-113_2eb73820An example of how the camera recognizes individual objects in an image. Note, for example, the detection of horizontal markings or a vertical traffic sign.

"The new EU regulation concerning the Intelligent Speed Assistance function explicitly states that, for three of the four types of tests required, results obtained under laboratory conditions may be used," explains Roman Václavík from the carmaker's Technical Development department. 

The performance of the assistance system is affected by a wide range of factors. In addition to current visibility conditions (day, night, fog, rain of varying intensity, snowfall, frost and so on), the results are also influenced by specific vehicle parameters, such as the windscreen specification (with or without heating), chassis settings (standard, lowered, adaptive and so on), or the type of headlights. "Each car therefore has to be tested in various combinations, which significantly increases the volume of testing," Václavík points out, highlighting the potential for savings in both time and costs. 

Intelligent Speed Assistance (ISA) 

ISA is now a mandatory safety system without which new vehicles cannot be registered for road use in the European Union. Using a camera and navigation data, the system monitors the current speed limit and warns the driver if it is exceeded. To obtain the so-called Certificate of Conformity (CoC), carmakers must demonstrate that this system functions adequately. 

Guarantee of Reliability 

For ISA homologation to be carried out using virtual testing, a special test rig had to be developed. It was also necessary to ensure that the simulations worked reliably and that their results corresponded with sufficient accuracy to how the assistance system would behave in real-world conditions. The independent technical testing and certification organisation TÜV SÜD Czech played an important role in assessing the reliability of virtual testing.

Navrh-bez-nazvu-117_0baaa250A test setup for virtual type approval (known as HIL) consists of specific components used in the vehicle and additional computer equipment.

"We helped translate the requirements for simulation reliability into the design of the entire simulation environment. This involved not only technical aspects, but above all determining what requirements the environment and the related processes had to meet so that their results could be used as reliable evidence for homologation," says Vladislav Kocián, a representative of TÜV SÜD. "We brought the perspective of an independent authority into the development process, as such an authority needs clear and verifiable evidence of the reliability of the results," Kocián explains. TÜV SÜD also monitored, for example, the transparency of the development of the simulation environment itself, as well as a number of other aspects related to virtual testing process. 

The result is not only a dedicated HIL (hardware-in-the-loop) test rig, but also documentation of approximately 400 pages. This describes in detail how the equipment meets the individual requirements for the so-called credibility of virtual test results and how the individual steps of the homologation process are carried out. This documentation forms the key basis for the patent application. 

Navrh-bez-nazvu-116_44157605During testing, the solution uses the camera exactly as it is installed in the production vehicle.

Just Like in the Real World 

The test rig includes a camera for traffic sign recognition, a GPS simulator, an antenna and an on-board infotainment control unit. Naturally, it also includes a computer that generates a virtual environment for the system under test, along with other vehicle equipment. "For example, we have a key reader here, because we found that physically implementing it is easier than simulating the presence of the key virtually," says Jakub Prchal from Digiteq Automotive, whose team worked on the solution. 

According to him, one particularly interesting element is the GPS simulator used in combination with the vehicle's standard antenna. The antenna is shielded from publicly available GPS signals, and the test rig sends it data about the simulated position on the track, generated by the GPS simulator. "This separation of the simulations ensures that the camera system responsible for recognising speed limits behaves as it would in real-world conditions," says Prchal. The routes used for the simulations are virtualised versions of real roads. This makes it possible to verify that the results correspond to those of real-world tests. At the same time, it is the only way to ensure that the ISA system can function in the laboratory at all, as it would not work without map data. 

Navrh-bez-nazvu-114_dd21fd88The team can also perform simpler tests by projecting the camera's image onto a screen in front of it. However, this procedure is not sufficient for type-approval tests.

What the Legislation Requires 

A total of four types of tests are mandatory for ISA approval, three of which the legislation allows to be carried out in laboratory conditions. The first test verifies the functionality of explicit traffic sign recognition itself, i.e. whether the camera can read a speed limit sign (or the sign indicating the end of the speed limit). The next required test checks whether the system provides a timely visual and audible warning when the speed limit is exceeded. It also verifies whether the warning stops once the vehicle's speed falls below the limit (the visual warning must stop, while the audible warning sounds only once during a single over speeding cycle). The third part concerns the recognition of implicit traffic signs, which indicate, for example, the type of road (motorway, road for motor vehicles, pedestrian zone and so on), and the system's response to the speed limits applicable in the country concerned. The reliability of the virtual tests is then verified by a fourth, physical test, during which an independent technical testing organisation randomly selects 400 kilometres of roads on which the aforementioned functions are tested under real-world conditions. 

The computer does not project image data onto a display for the camera to capture (although this method can be used for simpler tests). Instead, it sends the data in a special graphics format directly to the camera chip. "This ensures that the camera reads the environment as if it were real," says Prchal, explaining that projecting an image onto a display could cause unwanted deviations from real-world conditions. 
  
Incidentally, the virtual tests also run in real time, which is another requirement for ensuring that they correspond to reality. "They cannot be accelerated. The standard camera used in production vehicles would literally become confused," explains Prchal. Even so, the solution saves time, as tests can also be carried out outside normal working hours or over the weekend, and the time-consuming preparation of the vehicle is no longer necessary. There is also no need to find a driver. The first vehicles to use the new technology for ISA homologation in practice were the Peaq, Enyaq and Elroq electric vehicles in their current model-year versions. 

Navrh-bez-nazvu-115_1e086cc7The test setup also includes the instrument panel, the infotainment display, and the headliner, which incorporates certain functions that would otherwise be difficult to simulate during testing.

Future Development  

Škoda Auto and Digiteq Automotive intend to use their experience with the virtual homologation process to further develop computer-based testing. "Approval of the equipment by an independent authority is a sign of the reliability of the entire solution, which gives us a solid foundation to build on," says Martin Fischer, a representative of Škoda Auto's Development department. 

In addition to further homologation tests, the carmaker plans to use virtual simulations in the future, for example, to demonstrate the functionality of selected systems in Euro NCAP tests. It is also working on expanding the simulation capabilities to include additional sensors, not just cameras. "It is still a bit of a challenge," Fischer hints. The solution could also help improve cooperation with suppliers of cameras and other sensors.