Autonomous driving: A step-by-step reality or disillusion?
How can we overcome technology, data and cost challenges to make autonomous vehicles a reality? We considered the issues at the UBS Future-Now APAC Conference in Hong Kong in mid-June.
As of the start of this year, research by the McKinsey Center for Future Mobility forecast that autonomous driving could create US$300 billion to US$400 billion in revenue by 20351 . Yet whether it will fulfil this potential remains a source of debate within the industry.
Billions of dollars in investment have already been channelled in recent years into bringing the concept of ‘smart’, or intelligent, driving to life. However, from the US to Europe to Asia, the goal to transform the future of transportation continues to hit roadblocks such as supply bottlenecks that have pushed back timelines.
As a result, the vision back in 2017 that autonomous driving would be commonplace by 2022 hasn’t been fulfilled, according to Paul Gong, Head of China Autos Research at UBS, who moderated the discussion.
A key question today, is how to make level 4 (L4) cars – those that can achieve driverless control – profitable, asked Dr Ni Kai, Founder and Chief Executive Officer of HoloMatic Technology. “We need to understand how we can apply the technology commercially. It will be achieved through continuous iterations in the development of the technology.”
He is optimistic this will happen in the next three to five years, citing the quick pace of progress made by electric vehicles (EVs) in certain markets, including Mainland China.
Adolph Chiu, Vice President of Geely Automobile, agrees with the priority to resolve questions around how to scale the business and make it more efficient.
“The more expensive cars might be more powerful, but in all cases, costs must be controlled and we need to realise economies of scale across the range of available products.”
A roadmap for self-driving vehicles
A roadmap for self-driving vehicles
According to Gong, one of the pathways to autonomous driving is to take it step by step, involving an upgrade of the mass produced level 2 (L2) cars – those with partial driving automation – as well as moving their use from highways into cities.
Among the potential ways to scale the business in this way, added Ni, is to use L2 cars to bring more value to vehicle owners. For example, a selling point could be the use of the technology to help drivers when parking.
This reflects an increasing need within the industry to make the best use of existing technology, especially given the possibility of supply shortages of artificial intelligence (AI) chips, among some potential issues going forward.
“Cost is a core concern,” explained Ni. “Companies need to weigh the value of the hardware against the functions it needs to offer.”
Also important in making self-driving vehicles more widespread is ensuring that all cars, regardless of their level of luxury or price point, have the same safety features installed. “Whether cheap or expensive, autonomous cars need to have the same functions and parts,” said Chiu.
This might also lead to the required change in consumer mindset. For example, some people are wary that a driver is still needed for the majority of autonomous cars, since the AI requires a human to pay attention to the traffic on the roads. “Even though I am someone who is used to driving autonomous cars, I still feel quite nervous about this,” added Ni.
A data dilemma
A data dilemma
Ultimately, the use of data specific to each local market is a critical factor in effectively scaling the autonomous driving industry.
“A longer track record and the ability to access and interpret data are key issues in making autonomous driving a reality,” said Ni.
He said Geely is working on localizing such data across Mainland China, where Chiu believes the local dynamics enable much more data to be generated than in any other country. For example, the road system is more complex overall, with megacities, more cars and higher population density. “This enhances the machine learning capability to accelerate the development of autonomous cars,” he added.
At the same time, in response to a question from the session co-moderator, Thompson Wu, Head of A-share Technology Research at UBS, about the potential to share data, Chiu emphasized that protecting customer data is essential given the sensitive nature of personal information.
