The impact of Artificial\nIntelligence (AI) is being felt across a range of technologies disrupting the\ntraditional methods of doing things and pointing the way to futuristic advances\nin technology. Fadi Kanafani, Senior Director Middle East, NetApp, highlights the current\nstatus of AI in the automotive industry and considers a future where cars will connect\nto each other, to our homes and to infrastructure.<\/p>\n\n\n\n
The impact of Artificial Intelligence\n(AI) is being seen across industries and geographies. AI is now a key to\nsuccess for organisations and is set to be a significant contributor towards\nglobal economic growth by 2030. <\/p>\n\n\n\n
According to a study by McKinsey Global Institute (MGI<\/a>), on average, the global gross domestic product (GDP) could increase by\n1.2 percentage points per year, which would correspond to a total \u2018value added\u2019\nof US$ 13 trillion. <\/p>\n\n\n\n Among a number of industry\nsegments, the automotive industry is one of the most technologically advanced\nand progressive industries. It\u2019s no surprise that the industry is a frontrunner\nin adopting and incorporating AI into research, design and manufacturing\nprocesses for smarter and better outcomes and products.<\/p>\n\n\n\n When you think about AI in\nautomotive, autonomous vehicles is likely the first use case that comes to\nmind. While the holy grail in the industry is full self-driving, defined as\nlevel five, most companies are already offering increasingly sophisticated\nadaptive driver assistance systems (ADAS) as stepping stones towards that level\nof autonomy.<\/p>\n\n\n\n In an industry like automotive, the number of possible AI\nuse cases is large and essentially divided into four segments which are\nautonomous driving, connected vehicles, mobility as a service and smart\nmanufacturing. <\/p>\n\n\n\n Naturally, there are overlaps between some of these\nsegments; success in one area can yield benefits in another. For example,\nautonomous driving may be a key element of a mobility-as-a-service strategy.\nThere are also many requirements that all segments have in common, including\ninfrastructure integration, advanced data management, security, privacy and\ncompliance.<\/p>\n\n\n\n There are, however, challenges to\nachieving full self-driving. Each car deployed for R&D generates a mountain of\ndata; 1TB per hour per car is typical. Teams can\nexpect to accumulate hundreds of petabytes to exabytes of data as autonomous\ndriving projects progress. <\/p>\n\n\n\n This raises several critical\nquestions such as how to create a pipeline to move data efficiently from\nvehicles to train a neural network or how to efficiently prepare and label data\nfor neural network training. <\/p>\n\n\n\n Some questions that need to be\naddressed are how much storage and compute power is needed to train a neural\nnetwork, to run inference on a trained neural network and if the training\ncluster should be on-premises or in the cloud. It is also important to\ndetermine how to correctly size the infrastructure for data pipelines and\ntraining clusters including storage needs, network bandwidth and compute capacity.<\/p>\n\n\n\n Cars and other vehicles are quickly\ntransforming into connected devices, and there are a number of immediate use\ncases for AI in connected cars such as personal assistants\/voice-activated\noperations, telematics and predictive maintenance and infotainment\/recommenders.\n<\/p>\n\n\n\n Today, cars\nuse cellular and WiFi connections to upload and download entertainment,\nnavigation and operational data. In the near future, we\u2019ll also see cars\nconnecting to each other, to our homes, and to infrastructure. <\/p>\n\n\n\n For example,\nAudi has already introduced technology to connect cars to stoplight infrastructure<\/a>, enabling\ndrivers in select cities to catch a \u2018green wave\u2019, timing their drives to avoid\nred lights. That\u2019s just one of many\nopportunities to use data from connected cars. <\/p>\n\n\n\n