Elly Yates-Roberts |
The development of autonomous driving promises to transform the global automotive industry. However, the process to achieve completely driverless vehicles will be a long one.
“If you talk about autonomous driving, you’re really talking about the stretch goal which is full autonomy,” said Marek Jersak, vice president of autonomous drive at DXC Luxoft. “What we see now is most of the industry targeting Level 3 autonomy, which means providing automation under well-defined conditions like highway stop-and-go, combined with advanced Level 2+ driver assistance in other situations like highway cruise.”
According to Jersak, both automotive and technology companies need to change their approach to software development to succeed.
“For autonomous driving, you first need to train your algorithms, then validate them,” he says. “As the level of automation increases beyond Level 2+, so does that amount of training and validation data required – dramatically! This kind of data management at scale is not a core competency of automotive companies. We need processes that enable fast, data-driven development and continuous improvement, but you also cannot simply take the philosophy of mobile phone software and put it into a car. You need to adhere to strict automotive quality, safety and security standards. Building the bridge between the two approaches is key. This requires the automotive industry and the AI and data management sector to understand each other and bring together their knowledge.”
“Understanding how to efficiently manage and organise the data for the development of autonomous functions combing deep analytics, IT and automotive expertise is where we see our core value,” adds Matthias Bauhammer, Robotic Drive offering lead for data-driven development at DXC Luxoft. “We have understood the key autonomous driving use-cases and how to automate, integrate and scale the needed solution. We’ve done that by working with Automakers, Tier-1 suppliers and partners. We bring that wide understanding to the market and commit to one joint goal, creating a global, multi-tenant development eco-system.”
The Robotic Drive data-driven development platform is making use of Microsoft Azure to help the automotive industry change its development processes. “Partnerships are absolutely essential here,” says Jersak. “There is a huge number of technology building blocks which need to work together to make this development challenge manageable. We’re proud to work closely with Microsoft to run Robotic Drive on Azure. The advantages that we bring include giving end clients the flexibility to run their use cases and functions both on Azure and on-premises. We enable our clients to easily scale their data and compute usage up and down. It changes throughout the development lifecycle and with the type and number of parallel tasks taking place, and we can orchestrate all that with our deep understanding of automotive development.”
This article was originally published in the Winter 2020 issue of The Record. To get future issues delivered directly to your inbox, sign up for a free subscription.