Big Data Analytics Platform for ALP.Lab Proving Region
As part of a strategic collaboration, AVL List and NorCom Information Technology – teamed up to share their expertise and develop Big Data solutions specifically tailored to the automotive industry. By combining their respective strengths the ADAS/AD Data & Analytics Platform was established, which is now being used for the very first time at the Austrian proving region ALP.Lab.
The testing of automated driving functions requires the analysis of complex environment data. A state-of-the-art analysis platform must be capable of overcoming these and other challenges associated with Big Data. In AVL’s Data Intelligence portfolio, the software tool of choice used for this purpose is NorCom’s DaSense Platform. Of particular relevance as convenient programming interface is the Python extension: This tool abstracts the Big Data technologies needed for handling parallel calculation tasks and enables users to write simple sequential code. As a result, this ensures the scalability of analyses without the domain experts having to build up specialized Big Data know-how first.
To represent the environment model, the platform uses the object-oriented description language Open Simulation Interface (OSI), which reduces the complexity of environment model description and increases the platform’s compatibility with sensors from different manufacturers.
One frequently occurring task, which can now be executed with high scalability, is finding defined scenarios in a large quantity of data. If, for example, a vehicle exhibits irregularities in a certain situation, you might be particularly interested in analyzing how that vehicle performed previously in similar situations. Before this can be done though, the relevant scenarios must be found. However, considering the many driven kilometers of a large fleet, the task will seem like searching for the proverbial needle in the haystack.
Dr. Jost Bernasch, Managing Director ALP.Lab GmbH: “As an Austrian proving region for automated driving, ALP.Lab is provided with a modular, rule-driven workflow and analysis tool, which allows us to conduct comprehensive ADAS/AD tests and validations of driving functions on public roads and proving grounds. In doing so, extensive vehicle, sensor and infrastructure data are collected, merged, analyzed and further processed. Particularly when it comes to processing the huge volumes of data generated in testing, the AVL data platform with the integrated analysis tool DaSense is essential for handling the data bulks efficiently.”
In the ALP.Lab proving region AVL, Magna, Virtual Vehicle, Joanneum Research and the TU Graz pool their activities related to automated driving testing in Styria and across the border in Hungary. The proving region covers more than 400 kilometers of highways, public roads and closed proving grounds and is supported by the Federal Ministry of Transport, Infrastructure and Technology.