Systematic testing and validation of automated driving functions is one of the keys to bringing this technology to market maturity. In addition to the advantage of being closer to reality, road testing has the disadvantage that the time and cost required for complex automation levels for SAE Level 3+ is enormous. Currently, many institutions are working to largely replace road testing with X-in-the-loop methods ranging from virtual simulation (model-in-the-loop), component test benches (simulation, hardware, processor-in-the-loop) and whole vehicle test benches (human-in-the-loop, vehicle-in-the-loop). One of the greatest challenges is to reproduce the complex interaction between ADAS/AD sensor technology, vehicle guidance algorithms and vehicle actuators (drive, brake, steering) with sufficient realism on a test bench. The generation of relevant driving scenarios in scenario-based testing also turns out to be complex, since the criticality and relevance of scenarios are difficult to describe.
The project is concerned with the systematic testing of automated driving functions on a test bench concept developed specifically for the project. It enables overall vehicle integration and provides test capabilities for automated driving functions, the entire powertrain and highly dynamic driving maneuvers at the limit. For this purpose, an existing simulation environment is applied to a complete vehicle test bench with hard real-time in a first step. Available sensor models and sensor stimulators are implemented on the test bench and validated in detail with test drives under different environmental conditions. Driving on the already modeled in detail highway section of the A2 Graz-West to Laßnitzhöhe is chosen as scenario. A validated traffic flow simulation generates realistic behavior of vehicles of the surrounding traffic. These scenarios are extended with critical situations by so-called "stress testing" and determined representatively from an in-depth analysis of accident databases. The project concludes with an impact analysis of the presented concept, which demonstrates the efficient development of alternative powertrains, vehicle dynamics control systems and automated driving functions.
The innovation lies in the enormous closeness to reality which is achieved by real-time behavior of all components, digitalization of real routes, traffic flows and accident data as well as validation and verification of the system. Currently, there are no standards or laws governing the approval of automated vehicles, and the development of the test bench with the highest level of realism in the world also means the opportunity to play an influential role here.