Big Wheels – Big Data

Data reduces fuel consumption, enables precise forecasts on the vehicle, makes logistics more efficient and is absolutely necessary for automated driving. But how do you protect them and who do they actually belong to?

The sensors and control units in highly automated vehicles are already generating a lot of data. CGI: Fabian Techel

Nowadays, the average executive-class car typically features more than 100 control units. For a typical premium-class model, that number can easily climb to several hundred. Vehicle electronics generate up to a terabyte of data every hour – quantities that would completely fill the hard-drive or SSD of a standard laptop in only one or two hours. Industry consensus is that these numbers will continue to skyrocket in the coming years. Intel is working on the assumption that by 2020, sensors and control units of highly automated vehicles will generate data quantities of up to four terabytes per hour.

For reasons of capacity and cost, the majority of this data remains stored in the vehicle – mobile data networks are simply not yet capable of transferring multiple terabytes of data into the Cloud every hour. Nevertheless, modern vehicles are constantly conversing with servers and backend systems. In order to reduce the data for transmission into more bite-size pieces, the on-board electronics can aggregate the measurement values and sensor data into indexes, or extract only the values that are truly important. As a result, multiple terabytes are reduced to just a few megabytes of relevant data records per hour. The vast majority is simply deleted.

There is plenty of data being transferred in the opposite direction too – be it traffic information, map updates, or environmental information for the on-board systems. This of course excludes the comprehensive data exchange involved in logistics applications for the commercial vehicle industry. This can include constant updates on the position and interior temperature of containers as well as G-forces that may affect the cargo. All is transmitted to the logistics provider’s data centers, or those of the customer.

Connectivity reduces consumption

An important application area for digital networking is, for example, fuel efficiency and thus the reduction of CO2 emissions. While so-called efficiency assistants have only really featured in the most recent generation of passenger cars, they have been a staple in the commercial vehicle sector for a while. The extent to which Cloud-based data and on-board electronics can already interface is demonstrated by eHorizon, a commercial vehicle offering from original equipment provider Continental. “The sensors on eHorizon-enabled vehicles transmit their data directly to the Cloud,” explains Continental Spokesperson Sebastian Fillenberg: “Are there roadworks or narrowed lanes ahead? What is the temperature, wind speed and precipitation? Have the ABS sensors detected aquaplaning hazards or ice on the road surface?” This data is shared with other trucks from the same manufacturer, providing that they are also equipped to receive this information. This electronically-enabled vision of the road ahead allows the distance regulation system and automatic gearbox to adapt to the traffic and environmental conditions ahead, and ensures efficiency.

Big Data predicts malfunctions

This is of course just one of the many ways that vehicle data can be used intelligently. The top example is predictive maintenance – the ability to rectify problems before they arise. “It is our goal to give customers as much advance notice as possible for maintenance work and parts ­replacements, and thus make them easier to cater for in planning,” explains

Michael Kimmich, Manager of Mercedes Benz Trucks. He explains the concept – based on Big Data – using the example of a turbocharger turbine: “Using input values such as temperature and engine speed, we can perform a precise damage calculation and warn the customer before the turbo ultimately malfunctions.” Kimmich is confident that intelligent algorithms and ever-improving Big Data techniques will enable a very high prediction reliability. Ultimate accuracy is crucial for such concepts. Having the system “play it safe” and recommend a part to be replaced too early is not an option – the premature replacement of a working component would not be acceptable for customers. Of course, leaving the warning too late is equally unacceptable.

Vehicle manufacturers secure far-reaching access to the data accumulated in the vehicles. CGI: Fabian Techel

In-the-field data analysis helps developers

The networking of on-board computers has long been put to great use in vehicle development, as one solution from vehicle software business Elektrobit demonstrates: the business offers software that the vehicle developers can use to perform data analyses in subsections of a vehicle fleet, while they are in everyday operation. The system thereby allows the developer to identify all cars that have had trouble starting in the last 48 hours. In addition, it can instruct the responsible control unit to log the characteristic values of the battery. These records can then be anonymized and collected for evaluation. The results can then be used to improve both products and systems. A handy side effect is that the system is also able to proactively inform the driver of looming faults.

Especially the last example demonstrates why vehicle manufacturers go to such lengths in their terms and conditions to ensure far-reaching access to data generated from the vehicle. Data is the most important resource of the 21st century. The more business and revenue models are based on its usage, the more the question is asked “who does this data actually belong to?” Groups that demand unlimited access are numerous, and include vehicle manufacturers, system developers, mobile network providers, platform developers such as Google, insurers, leasing firms, the public sector to name but a few. Expert organizations will also need access to relevant information for vehicle inspections.

Privacy rights versus economic interests

While purely vehicle-centric data is comparably noncritical in terms of privacy rights, the same cannot be said of driver data, such as the driving- and rest period information generated by digital tachographs. “While the last few years have been characterized by somewhat of a Wild-West mentality surrounding data – everybody just took what they wanted, using whatever legal tricks they could – new judicial frameworks such as the General Data Protection Regulation may lead to fundamental change,” posits Andrea A. Voßhoff, who was named Germany’s Federal Commissioner for Data Protection and Freedom of Information in 2014. There is increasing demand for neutral authorities that will ensure interests are balanced when handling data. Similarly important is the security of communication paths and data-processing systems. “The driving force for cybercriminals is material gain,” explains Ruben Lirio, a Specialist for DEKRA’s Cyber Security Evaluation Service. “The more money that can be made with the data, the higher the motivation to steal it, use it to compromise others, attack the value creation chain and attempt extortion.” Data security and a fair, yet not overly bureaucratic balancing of interests in data usage are therefore among the most important tasks to address in the booming data economy. Lirio maintains that DEKRA is well-equipped to tackle both.