Data analysis for tomorrow’s mobility

In addition to electric drive energy, there is a second important fuel for the driving experience of the future for modern vehicles: Data. More and more digital functions will soon turn the car into a smartphone on four wheels – with many new services for consumers and business opportunities for manufacturers. Massive amounts of data must be measured, modeled and analyzed. The collection and storage of data is also an important factor in launching the mobility of the future.

With digitalization, data volumes around mobility have been increasing for years – and the trend continues to rise. Automotive engineering service providers use big data methods for fleet analysis. To draw conclusions about the perception of the environment, information is collected via sensors, lidar, radar or camera systems, for example, and evaluated in the shortest possible time. It depends on specific data know-how and the ability to translate customer requirements into knowledge, as well as on fast, precise, and secure technologies for processing the data.

Data Management: Prepare heterogeneous data for analysis

More and more sensors, processors and other digital technologies are now being used in vehicles. It requires processing and controlling vast amounts of data from a variety of sources. To achieve this, a reliable IT infrastructure is needed. The keys to this are large, high-performance storage systems, high-performance and scalable analysis systems, and efficient and secure data management.

But how is the process going in detail? Even before data recording can take place, there are many preliminary tasks to be completed: First, the investigation objectives and recording concepts as well as a test planning must be worked out in accordance with the usage scenarios. Test vehicles must be set up or converted and the measurement technology integrated with the appropriate configuration.

A particular challenge here is that the data collected comes from different data sources such as cameras, route paths or driver profiles and therefore usually have different formats. However, these heterogeneous data must be linked and analyzed together, because it is only through the combination of the numerous signals that it is possible to recognize complex error cases or states. For example, the causes of engine start-ups can be identified, or the comfort functions can be automatically adapted to the driver.

Data analysis: Translate customer requirements into insights

For the analysis, the data is transferred via the appropriate IT architecture to a measurement data management system. The web-based measurement data management system of IAV, for example, enables a secure and structured administration of vehicles, measuring devices and the resulting measurement data: Vehicle developers can search through the enormous data volumes based on various criteria and quickly access the required data records. The measurement data are kept for any period of time, so that entire vehicle fleets can be examined over several years.

The analyzes based on the measured data are carried out according to customer requirements or according to freely definable criteria. They provide the customer with an automated result report from which appropriate recommendations for action can be derived. Customers can also access intelligent decision-making aids for their operational and strategic plans on an interactive dashboard.

The expertise of data engineers such as IAV’s lies primarily in translating customer requirements into measurement results. Together with the customer, they work out which question is to be answered exactly, which measurements are necessary for this, and which data are to be calculated against each other in order to provide the corresponding proof. On this basis, the IAV then develops algorithms that are able to provide accurate answers to the questions.

Powerful storage makes analysis easier

To deliver these services reliably and efficiently, analytics must be backed by a reliable and powerful backend. The demands on the infrastructure in terms of compute, network and storage performance are high. IAV is based on IBM Cloud Object Storage and IBM Spectrum Scale. This choice provides the IAV with valuable advantages when working with the data:

  • Security: The storage can encrypt data on the fly to meet customer security and compliance requirements.
  • Precision: Within the encrypted data, a certain storage position can be exactly controlled and processed without having to touch the entire data packet. This also saves money because there is no additional investment in RAM and CPU.
  • Process: This exact control of the stored data saves a lot of time, in which complete data packets would have to be copied, unpacked, the desired area extracted, and the package would have to be tied up again.
  • Efficiency: Deleting data on the storage system is the same as deleting data on a hard disk; it is removed from the list when it is deleted and is therefore no longer visible.
  • Speed: IBM Object Storage provides fast storage operations for large amounts of data. It is the Porsche among the heavy stores.

Growing data volumes require hybrid solutions

The automotive industry is also increasingly relying on cloud solutions for its diverse requirements. It enables flexible scaling of resources, especially at peak loads. The question that is becoming increasingly apparent is what a high-performance infrastructure looks like in which the cloud and on-prem function as a hybrid model. One thing is clear: The more we focus on fully autonomous mobility solutions, the larger the data volume that storage must handle. To deliver accurate, fast, and reliable results, data experts like the IAV need systems and networks that can map the growing demand for performance. Whether in the cloud, on-prem or hybrid.