High-Performance Analysis for Smart Decisions

Collecting and evaluating large volumes of data is an important part of automotive engineering. Increasing digitalization makes it indispensable. But only a few OEMs are aware of the extent of big data analytics and in a position to make effective use of the resulting potential. Here IAV offers a holistic big data value chain beyond the conventional recording and analysis of measured data. IAV’s measurement data management with interactive analytics dashboards provide intelligent decision-making tools for operative and strategic projects for operative and strategic projects.

Complex connectivity, growing automation and increasing digitalization are producing more and more data which then have to be collected and analyzed. One central challenge in automotive engineering will therefore consist in integrating these quantities of data in the business processes in order to create value and make efficient use of the resulting potential.

OEMs often use restricted measurement recording with manual fault analysis processes, thus forsaking extensive know-how about vehicle status and test drives. The significance of automated validation mechanisms is growing all the time, particularly when it comes to future topics such as autonomous driving.

IAV can draw on more than 15 years of experience in engineering, validation processes and in supporting the product development process in the automotive industry. Large data volumes are part of day-to-day work. “Some test drives generate several terabytes of data each month”, reports Björn Steffen, Senior Vice President for Telematics and Products at IAV.

IAV’s end-to-end tool chain covers all relevant activities, such as constant recording and automated data transfer, as well as storing, managing and analyzing measured data. The focus is on using the generated data to create value as well as deriving correlations and patterns. Here IAV uses the latest big data technology, giving priority to top-quality and high-speed data analysis, also in real time.

Besides collection and analysis, the evaluation of data and information is crucial for providing the basis to take operative and strategic decisions. IAV has developed a big data analytics dashboard that converts huge data quantities into usable information to provide a comprehensive overview of all vehicle data. The interactive dashboard has various diagrams and widgets that visualize important parameters and use cases. The diagrams include thermal maps, histograms and signal profiles. “We have also implemented a broad selection of widgets, featuring a video player, map with GPS tracking and the depiction of diagnosis extracts”, adds Jens Schulze, head of the Fleet Validation and Automotive Data Analytics department at IAV.

The dashboard has a user-friendly interface for easy and individual configuration of widgets and diagrams according to user preferences without needing any programming skills. The overall data for the vehicle or fleet can then be used to produce detailed views for studying relevant processes and use cases. This makes it faster and easier to visualize workflows and anomalies for error detection. The user can look at the measurements in retrospect with the replay. Direct live streaming of the vehicle data is also possible while the vehicle is still moving.

Moreover, the data management system developed by IAV gives the user the option of complex searches across all measurements with a special search user interface. The results are visualized automatically within seconds, highlighting interrelationships and outliers. Various meta data can be added to reveal correlations and relationships between the vehicle signals that may possibly be hidden in older data records. For example, the analysis are combined with GPS data to put the results in a geographic context. In combination with the analytics dashboard, it is possible then to take a detailed look at inconsistent measurements for iterative narrowing down of errors.

Behind the visually processed data on the dashboards is a comprehensive measurement data managament system and a high-performance analysis framwork. IAV’s web-based data management system is a proprietary software development that offers numerous features for managing and organizing measurement technology, vehicles and fleets. Other possibilities include defining automated analytics and downloading visualized reports about the evaluated results.

The measurements are evaluated with the IAV analytics framework that processes a large variety of data formats. The use of object storage and parallel virtual machines distribute the computing load across hundreds of nodes so that even real-time evaluation of data is possible. New technologies result in an open, scalable architecture that includes the possibility of connecting various systems. Depending on the customer’s specific requirements, other external analysis tools can also be used alongside the IAV analysis framework.

Regardless whether the focus is on cold climate testing in Scandinavia or test drives in South Africa, IAV’s flexible, open system for managing measurement data can be used for distributed data analysis all over the world without losing performance or safety. To this end, IAV uses its own local computer clusters at the specific location for data storage and analysis, al-though the main data center is still based in Germany. That means that only the analysis results have to be transferred and not the huge quantities of raw data. The whole system can even run in a private customer or in a public cloud. The vehicle developer can thus monitor all test drives across the globe in real time from his local workstation with the possibility of intervening directly in the event of any errors.

Test and quality management in particular offers many use cases where IAV’s all-encompassing analysis value chain offers great potential for optimization. Developing and testing new components and various vehicle applications is a time-consuming, costly process. Fast availability of test results not only reduces costs but also offers a competitive advantage by speeding up product launches. ”The use of real-time data processing and predictive analytics mean that tiny anomalies and errors in the test process are revealed immediately“, says Schulze. Early intervention is thus possible with shorter test cycles. In turn, this reduces time and costs involved in the testing process and in analyzing the data, thus shortening the whole vehicle development phase.

There are already plans for further developments with the agenda including pattern recognition across all measurements as well as the use of artificial intelligence. ”In this way we are technologically highly innovative in our field and are prepared for future topics as well as future challenges.“, adds Steffen.