Going all the way, we deal with safe driving functions, from start to finish. And we do that for so-called functional prototypes, or for the hundreds of thousands of series production embedded systems in various vehicle platforms. To put it in a nutshell, we…
We use our “safe driving” functions and algorithms to predict potential danger situations at an early stage and to prevent a normal driving situation from turning into a threatening accident situation. Drivers receive information regarding danger spots on their route and in their surroundings and are assisted through various driving maneuvers such as lane changing or turning. External sensors and swarm information are utilized to represent and interpret the driving environment.
Our “accident avoidance” solution detects traffic situations that would develop into an accident without actively intervening in the vehicle dynamics. By using crash probability and timeframe calculations for example, avoidable vehicle movements can be activated in the form of automatic braking or steering intervention, thus making it possible to prevent accidents or reduce their severity.
In the area of “protection” we develop functions, methods and tools to provide the best possible protection to vehicle occupants and other traffic participants in case of an accident. These range from accident severity prognoses and predicting the repercussions of an accident with the help of human models based on finite element methods (FEM), to developing and assessing the effectiveness of novel, intelligent protection systems. This in turn is based on activities such as data analyses/ data science from our accident research, which we rely on to create an understanding of how accidents originate and the injury mechanisms at work.
The area of “care afterwards” revolves around securing the accident site and the optimal rescue chain. Because it can come down to a matter of seconds, we provide as much relevant data as possible when an accident occurs. This can be helpful for emergency services personnel who receive an automated emergency call with information such as the location and severity of the accident and number of occupants. With Car2X technology and connected functions, we create messages warning other traffic participants of the dangerous situation on their route. Our accident research team records the accident and uses the information to supply important findings for further optimizations and if appropriate, new products.
We link vehicle data with information from ambient sensors to anticipate dangers at an early stage and activate protection strategies, such as in the event of an imminent, unavoidable collision. We possess years of in-depth expertise in the field of automotive environment detection and work with current sensor technologies such as radar, LiDAR (light detection and ranging) and cameras for data acquisition and object analysis.
In the area of algorithm and software development, we rely on strong and uniform methods and processes. We also place high value on the early use of simulation technologies.
Our algorithms and predictors are developed in line with top-down design principles for detailed system understanding, then implemented and validated in compliance with the ASIL (automotive safety integrity level) classification scheme. Within this context, we work with innovative tools such as the SceneInspector for incremental probabilistic simulation or with proven frameworks such as ADTF (automotive data and time-triggered framework), that continuously support the system engineering process.
Functions, systems and software for integral safety must be developed in accordance with international automotive standards. Internal and external audits, as well as our projects, regularly confirm this compliance such as with respect to the ASPICE (automotive software process improvement and capability determination) level 2 requirements.
We utilize virtual development methods such as finite element simulation (FEM), stochastic simulation and driving simulation. With these methods we can generate the basis of data for the development and design of our integral driving and safety functions at an early stage, thus significantly increasing the efficiency, quality and field of effectiveness of our functions.
Our VAIVA team covers the entire chain, from driving simulation to FEM crash simulation and the enhancement of human models. This enables us to draw the correct (and important) conclusions with respect to component and structure design for the maximum level of vehicle passenger safety and other traffic participants.
When we develop electronic components used in semi-autonomous and autonomous driving and safety functions, the work obviously doesn’t stop there. We also have to carry out testing and validation. We love testing, an area where we incorporate intensive expertise along continually refined processes as well. And of course we fulfill industry-specific requirements such as the ISO 265262 standards for functional safety.
We validate individual components, as well as connected electronic control units or complex vehicle functions in line with SiL (software-in-the-loop), HiL (hardware-in-the-loop) or integrally (XiL). This encompasses the design, automation and management of complex component test beds for electrical, electronic and functions-oriented tests, or simulation and modeling by means of complex HiL models in MATLAB or Simulink.
We develop various customer functions within the framework of the 360° protection feature. For predictive front protection, automatic braking addresses various scenarios in parallel and cross traffic situations. Rear surveillance enables warnings when vehicles are stopped or parked, but also during lane changes. The local hazard warning warns the driver of danger spots and dangerous situations at an early stage on the basis of V2X direct communication with other road users, while proactive occupant protection only intervenes in accident situations. The interaction of potential crash prognoses between active and passive safety rounds out the portfolio.
Does the function really work like it is designed to? Or do we need to make a small change here or there? To find out, we utilize the virtual test drive system. This project combines the virtual world, the virtual vehicle and virtual people in order to influence the development of driver assistance systems at a very early stage. That way we help to make the development of tomorrow’s systems more efficient and faster.
APM is a project with results that are employed in many areas. Relying on our experience in the field of active and passive safety and autonomous driving, we continuously develop the process model that ensures our project activities comply with ASPICE. We then use this as a foundation to provide consulting services and put our customers within the VW Group in a position to design their processes in accordance with norms and by taking into account specific requirements such as functional safety. Apart from creating and improving processes, we also find suitable methods in order to guarantee the best possible project results.
IDE self-service is an internal project in which we developed something that has been close to our heart for some time now: a remote development infrastructure that allows you to work in an isolated, reproducible development environment. Our editor of choice is VSCode, a platform that offers two advantages. First, it provides access to project-specific tool chains within a matter of minutes, allowing new employees to get completely up-to-speed in a short period of time. We also use the platform to directly share tool chains and complex configurations and keep them up-to-date, regardless if C, C++, Python or Go is utilized.
The front camera serves as the basis sensor for a wealth of driver assistance functions such as automatic emergency braking, lane change assistants or traffic sign detection. The broad range of functions demands a high level of quality and performance, especially highly-automated driving, which calls for extensive testing depth and quality analyses. To ensure this happens, we continually test the image processing aspects during the development phase. This involves carrying out customer acceptance testing of the image processing software and the corresponding perception modules (lane/ vehicle/ pedestrian detection…), which allows us to issue release recommendations or rectify any problems.
Are you ready and eager to help us solve the really difficult technological issues?
