Carmakers suggest cross-vendor real-time cloud
Cars are equipped with an increasing amount of sensors, and these sensors generate real-time data. Most of the data are processed and used in the vehicle itself for purposes like engine control or driver assistant systems. With the vehicles being increasingly connected and more and more functions being implemented in some kind of backend system, in the industry has started a discussion about the value of these data and further use. For example they can be aggregated and processed to generate a real-time street map, “living” street map that on the fly can identify temporary obstacles such as roadblocks, traffic stalls, accidents and the like.
At the recent Electronics In Vehicles (ELIV) congress in Baden-Baden where the top designers and software developers of the European automotive industry met, the issue of the value and use of automotive data was played by Konrad Hübner, project manager of Connected Driver Assistance Systems and Cloud Platform for BMW. “There data are incredible valuable”, Hübner said. The reason: They could be the basis for optimised, seamless and individualised mobility services. An example is camera-based traffic sign recognition: Today, the findings of such systems are displayed at the dashboard to the driver of the respective car – and only there. But they could be fed into a living electronic map in real-time, Hübner suggested. Likewise, temperature, weather, road conditions, traffic congestions, deviations – all these things that annoy today’s motorists because their Navigation system has not shown they could become available via the cloud to every driver.
To give an impression about the amount of data: Hübner estimates that BMWs equipped with traffic sign recognition will have identified meaning and exact location of some 4 billion traffic signs by the end of the coming year. And these are the traffic sign data alone. With all relevant sensors together, a fleet of just 5 million connected cars will generate significantly more data than, for example, are created every day on social network platform WhatsApp with some 700 million users, Hübner calculated. This results in the need of a giant data processing platform in the cloud – too big for any single carmaker to fund, establish and run. The technological aspect would relatively easy to handle, Hübner said. “We do not need to reinvent the wheel. And not Super Big Data is our core competency, but designing excellent vehicles. But there are many suitable offerings at the market.” In any case, it would make sense for the makers and the owners of the vehicles to share these data – this way, the critical amount for a really meaningful and complete living map would be much faster to achieve. What’s mote, the living map would the basis for a wealth of new location-based services.
Therefore, a cooperation across the automotive industry would be essential, the top designer concluded. “We need standardized interfaces and high-quality sensor data”, Hübner said. The cloud platform could run applications based on geodata, analytics, and common services. In such en environment, a killer application like real-time traffic forecast would be just a commodity; a host of individualised applications could run on top. But in would require a joint effort. “This will work if all collaborate”, Hübner said.
BMW’s Hübner was not alone with his vision. Amer Aijaz, Director Strategy and Concepts for Electrics, Electronics and Propulsion of Volvo, happened to be the next speaker at the conference. “To reduce the number of traffic fatalities, we need automated driving”, he said. And for automated driving, we need a living street map with data processing in the cloud,” he said. “No OEM can do this alone. Volvo will definitively participate in such a project.”
Standardisation of vehicle data makes progress