Microsoft tackles the connected car

January 03, 2017 // By Christoph Hammerschmidt
For a long time, software giant Microsoft has been notably absent from the growing market around the connected car. Apparently, this is about to change. At the Consumer Electronics Show (CES) in Las Vegas, the company showcases its vision of connected driving. Members of Microsoft’s ecosystem are, among others, chipmaker NXP and Berlin-based automotive engineering company IAV.

Microsoft’s platform of choice for connected mobility solutions is Azure, a cloud infrastructure providing functions to create virtual machines, apps, websites, platforms-as-a-service, data management services and much more. At CES, Microsoft is highlighting scenarios where Artificial Intelligence bots are used to improve driver safety or integrate personalized functions. In addition, Azure-based applications are shown that acquire and analyze real-time data regarding the traffic situation including pedestrian density.

As eyes and ears, the Microsoft system uses up-to date automotive sensors like radar, lidar, cameras and V2X. Here we encounter NXP: The chipmaker is providing its V2X technology for the Microsoft solutions. At the Las Vegas show, NXP shows V2X road safety and traffic flow solutions that function as data sources for the Microsoft cloud platform. Use cases include collision warnings, smart traffic light and pedestrian detection solutions. These solutions are based on the RoadLink technology jointly developed by NXP and Cohda Wireless.

In an interview with EE Times Europe, NXP CTO Jens Reger detailed the scenario of vehicles that share data with others and are itself an integral part of the Internet of Things: At the car level, NXPs Bluebox real-time computing platform is extracting relevant data from variables like vehicle position, speed, engine and environmental parameters as well as sensor data and forwards these data to the cloud. Here, Azure-based AI applications evaluate the data to detect relevant patterns that can be used for many different types of applications. “The car collects the impressions it has gathered throughout its driving day and uploads them to the cloud”, Reger said. “There, it is merged with the experiences from other cars to form a superset of data, a kind of collective memory containing experiences with difficult situations and allows the vehicles to share their experience and to learn together. “This allows technology vendors to optimize their algorithms for automated driving”, Reger explained. This kind of “off-board learning” can benefit from the virtually