{"id":7488,"date":"2019-05-17T08:14:44","date_gmt":"2019-05-17T06:14:44","guid":{"rendered":"https:\/\/www.microsoft.com\/de-de\/industry\/blog\/?p=7488"},"modified":"2019-05-17T08:14:44","modified_gmt":"2019-05-17T06:14:44","slug":"a-good-simulation-ensures-real-world-driving-safety","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/de-de\/industry\/blog\/automotive\/2019\/05\/17\/a-good-simulation-ensures-real-world-driving-safety\/","title":{"rendered":"A Good Simulation Ensures Real-World Driving Safety"},"content":{"rendered":"
The term \u201ctest track\u201d is synonymous with the concept of automobile development. On the road to market launch, pre-production vehicles sometimes drive millions of kilometres on test tracks. If companies tested each and every improvement to the software required for autonomous driving, the amount of time and cost involved would be astronomical. The solution? Simulations in the cloud.<\/p>\n
<\/p>\n
<\/p>\n
<\/p>\n
When processes typically used in software development come face to face with other aspects of business, there can sometimes be friction. In the software world, short development cycles, regular updates with improvements and bug fixes are commonplace. But these mechanisms are foreign to the world of automobile development: If a new model makes it to the dealer\u2019s showroom floor, it has already passed a thorough round of testing. Its functional scope will not go through any more changes until the next facelift. Audi uses Microsoft Azure to create a simulation environment that draws on a massive pool of real raw sensor data. The information is gathered from more than a half million intersections, over 400,000 traffic lights, just under 320,000 bridges and more than 60,000 speed bumps by test vehicles worldwide. During a simulation, the software for a control device and sensor (system under test) runs in a virtual environment and processes the raw data generated by the sensor. If the control device\u2019s software responsible for vehicle safety makes a poor decision during the simulation, the consequences pose no danger and the error is rectified in the next version. Audi uses Microsoft Azure to create a simulation environment that draws on a massive pool of real raw sensor data. The information is gathered from more than a half million intersections, over 400,000 traffic lights, just under 320,000 bridges and more than 60,000 speed bumps by test vehicles worldwide. <\/p>\n","protected":false},"author":127,"featured_media":7491,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ep_exclude_from_search":false,"_classifai_error":"","footnotes":""},"categories":[208],"post_tag":[577],"content-type":[283,232],"coauthors":[67],"class_list":["post-7488","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-automotive","tag-azure","content-type-cloud","content-type-kundenreferenz"],"yoast_head":"\n
\nBut now, software is a requirement for bringing autonomous vehicles to the streets. It processes the data provided by various sensors, uses them to gain an understanding of the driving situation and then decides what to do next: accelerate, break, change lanes or turn at the intersection. And there is no room for error in any given driving situation. For example, the software must be able to detect pedestrians regardless of the light, weather or traffic conditions.
\nBut how can this critical software be perfected during the development phase in the most economical way? Take Audi AG\u2019s answer, for example<\/a><\/strong>: \u201cIn addition to real-world tests, we rely more and more on simulations. It\u2019s simply not practical to test each new software version for ten thousand hours on various continents,\u201d explains Adrian James, responsible for automated driving safety at AUDI AG. <\/p>\nEconomically viable simulation is possible only with the cloud<\/h2>\n
\nThe cloud is critical for this amount of data. Experts believe that next-generation vehicles will generate four gigabytes of raw sensor data per second. To be able to run the same number of simulations, approximately 200 petabytes of storage capacity and the same amount of computing capacity is needed. This is not economically feasible for a company whose core business is developing state-of-the-art vehicles, not operating a data centre.
\nAnd Azure customers need not worry about data protection, as Dr Stefan Gangl knows. He is a design engineer at Audi and explains, \u201cAzure behaves just like our very own data centre. It even manages the cryptographic keys we use to protect our intellectual property.\u201d Ultimately, the cloud brings the various developmental speeds in line with one another and unites the automotive industry\u2019s intrinsic rigour with the dynamic nature of software development.<\/p>\n","protected":false},"excerpt":{"rendered":"