INNERVATE

Germany is one of the largest automotive industry countries in the world. Several of the world's leading vehicle manufacturers are based there. The current vehicle development process includes, among other things, the evaluation of driving behaviour, in which a test driver performs a series of standardized driving manoeuvres using special measuring technology. The driving dynamics characteristics recorded during these tests allow the experts to analyze and evaluate both driving safety and comfort for the driver and passengers in various extreme situations.

Driving dynamics tests are an important and necessary component for modern vehicle technologies such as self-driving cars and electric vehicles, as they enable the automotive industry to bring safe and therefore commercially viable vehicles onto the market. The entire driving dynamics test process comprises three main phases. It begins with the installation and commissioning of the sensors and data acquisition in a vehicle, followed by the execution and evaluation of a single maneuver on the test track and ends with a comprehensive analysis of the collected measurements.

In this project, we will particularly focus on the second phase: the quality control of the maneuver execution. During the execution of the maneuver, a test driver must perform a series of measurement runs (i.e. drive the same maneuver or the same test several times) in order to collect a sufficient number of measurements (i.e. measurement series) for the later evaluation of the vehicle driving dynamics. To ensure an accurate evaluation, the collected measurements should be complete, i.e. cover all required ranges and provide sufficient measured values. In other words, each driving maneuver should be performed by high-quality driving according to predefined criteria.

The concrete objective of INNERVATE is to develop a solution to automatically evaluate the quality of driver maneuver execution and to support drivers with information about the parts of the maneuver where they have made mistakes. This involves the evaluation of measurement series, finding errors, and explaining what needs to be done to avoid these errors. In addition, it should be possible to generate valid measurement series from partially correctly executed maneuvers.