No contradiction: sheer driving pleasure and autonomous driving are indeed compatible, says BMW’s Dr Ian Robertson. By Martin Kahl
‘Sheer driving pleasure’… ‘The ultimate driving machine’… BMW is a company bullish about the very act of driving. At the heart of everything BMW does, everything it develops, produces and promotes, is driving, and the driver. Yet BMW is also at the forefront of driverless car development.
BMW board member Dr Ian Robertson speaks openly about his love for driving – and for driving fast. He is also a keen proponent of the hands-free car. As the Member of the Board of Management of BMW AG, Sales and Brand BMW, Aftersales BMW Group, to give him his full title, told Megatrends, manual driving and automated driving are highly compatible.
The OEM talks not of semi-autonomous and fully-autonomous driving, but of highly and fully automated driving, and although the company’s long-term planning includes cars that require no driver, the focus is on developing technology that enables a driver in the driving seat to take time off when the going gets tedious.
According to the five levels of autonomy as defined by the Society of Automotive Engineers (SAE), Level 3 is the first stage of hands off the wheel, and eyes off the road. Even at Level 4, fully automated driving, there needs to be a driver in the driver’s seat. “At Level 5, the driver is not required,” notes Robertson. “We are pushing towards Level 4. We’re in Level 3 at the moment, and the length of time during which you can take your hands off the wheel and eyes off the road is extending, but we’re very conscious of the limitations of the technology before we enable Level 4.”
Indeed, he believes Level 3 is “probably not possible under all circumstances.” Some companies, he says, are aiming for Level 4, bypassing Level 3. “Our view is that you need to develop the technology through Level 3 to get to Level 4.”
Driver no driver
Level 5 technology, then, appears to sit rather awkwardly with BMW’s whole ethos of driving pleasure. Robertson, however, sees no contradiction. Quite the opposite: “The two fit very easily together. We are renowned for Sheer Driving Pleasure, and the Ultimate Driving Machine. The word driving is very much in our DNA. And it’s very much in the brand positioning that our customers enjoy. But, of course, we live in a world today where congestion is more and more of a challenge, particularly in the urban environment, but also on the highways. And there is more than enough opportunity to engage the technology so that when you can’t have that sheer driving pleasure, you are able to do something else. So, yes, I think the two fit very easily together.”
The word driving is very much in our DNA. But congestion is more and more of a challenge, so when you can’t have that sheer driving pleasure, you are able to do something else. So, yes, I think the two fit very easily together
Robertson cites the example of his own daily commute into Munich from 35 kilometres (21 miles) south of the city. The first 25km, he says, involves Autobahn driving with unrestricted speed – some drivers frequently drive at around 250kph (155 mph). “But the last 10km or so involves driving on the Mittlerer Ring, the Munich arterial, and it’s bumper to bumper traffic that varies between zero and 30kph, and then back to standstill. It’s here that the semi-autonomous technology works really well. I can take my eyes off the road, my hands off the steering wheel, and I can check my email. At that time, I’m just effectively in a train of traffic.”
The transition from Level 3 to Level 4 will be two-fold, says Robertson. “The first is the technological development, the second is the legislative framework that will enable this type of technology in certain circumstances but not all circumstances.”
The most likely application of Level 4 technology is where traffic is controlled, such as on Autobahns and motorways, he notes. “You’re effectively segregated from oncoming traffic by barriers, and you have lane discipline which makes it easier to control. If you move into highly urban environments where there are so many variables, the technology development has to be much stronger.” Robertson expects the technology for both of these applications to be developed in parallel.
The handover time is a crucial area of debate. “The conscious time that drivers are alert to the road condition is dependent on the amount of time they are allowed to take their hands off the wheel and eyes off the road,” he says. “We are restricting that technology because we know that if you go beyond a minute, or a minute and a half, the distraction is much greater and therefore, the ability to take control of the vehicle, should you need to, is more challenging.” The question is how and when self-driving technology can and cannot be activated.
“The last 10 kilometres of my journey that I mentioned takes, depending on the traffic, between ten and 25 minutes. Because the Mittlerer Ring is well controlled with segregated traffic, it’s quite easy for our technology right now to enable me to take my hands off the wheel and my eyes off the road for that stretch of the journey. It can be done, and we know how to do it. But we also know that if I do that, and something happens, then the ability to come quickly back into the driving position is not as easy.” For this reason, BMW, like several other vehicle manufacturers pursuing autonomous drive technology, prompts the driver to take back control by, for example, putting their hands on the steering wheel. “It can be as little as a touch,” notes Robertson, “but it’s definitely a conscious decision to acknowledge that you’re still responsible for the vehicle.”
One man’s boring…
In 2014, BMW demonstrated its autonomous drive technology at CES in Las Vegas. There, the OEM emphasised the use of the technology not for slow speed local city driving, where bumper to bumper and start-stop traffic is the tedious, frustrating and increasingly common aspect of driving, but instead for what it said were long stretches of boring highway driving. The driving pleasure that a driver might get out of their vehicle discussed then would come from winding country lanes, mountain roads and picturesque landscape. Clearly, what some people enjoy, others do not – a conundrum for those developing autonomous drive technology. We put this to Robertson.
“I see different applications. I like to drive at 250 kilometres an hour on a highway. But the thrill – and this is one of the thrills of living in Germany – is that at that speed you are completely alert and you are very conscious of the excitement that the speed brings. For some people, that’s not what they do. They get on the highway and they drive at 120 kph, and they don’t overtake. For them, that could be the moment to switch on the autonomous drive technology.”
From the point of view of what you can do as a passenger in an autonomous car, the flexibility that many people were thinking about is not possible because motion sickness will prevent it
Robertson points out that BMW’s automatic cruise control system is activated after 210 kph, adding a level of security with automatic braking. “On the other hand, in the downtown environment where it is bumper to bumper, there isn’t much for you to enjoy. Of course you can enjoy the other features of your car, but you’re not going to enjoy sheer driving excitement. So one person’s excitement is another person’s tedious drive. I think the technology applications will suit different people in different circumstances.”
From HERE to there
Bringing autonomous cars to market is about much more than developing the vehicle technology. A range of stakeholders are involved, from infrastructure to telecommunications, and many are non-automotive players.
A cornerstone of BMW’s autonomous drive programme, says Robertson, was the acquisition of HERE from Nokia, in which BMW Group invested along with Daimler and Audi. The OEMs have since been joined by Intel and a number of other companies. “We realised very early that the digital map was the framework foundation for autonomous driving. This doesn’t mean we require a significant processing capability in the background. We will try to do most of this in-car, but we obviously need a big back-end, artificial intelligence tool.”
Robertson points to BMW’s partnerships with Mobileye, an Israeli start-up which was not originally an automotive supplier, and Intel. “Intel was not known in automotive either, and it has become a much more embracing supplier in this regard. Mobileye then was acquired by Intel, which tells you that the partnership selection was a good one.”
In December 2016, Tencent bought into HERE, followed a month later by Intel. Delphi is working closely with HERE, and at the time of writing, Continental and Ford are both thought to be preparing similar announcements about joining HERE. “And although I won’t name names, there are many others now saying they want to be part of this partnership, because they can see that this technology, and the way we’re structuring it, is likely to be one of the industry standards, if not the industry standard.” Robertson emphasises that the partnership which is developing will not be unique to BMW – but this is its strength, he says: “We fundamentally believe that the standardisation of this technology will be one of the things that the regulators want. Because the regulators, of course, want to see vast amounts of testing before they allow or enable.”
It is widely recognised that testing and validation are essential for the successful implementation of autonomous drive technology. Varying cumulative mile or kilometre distances are cited by different OEMs and technology companies as the minimum required for validating autonomous drive technology. At BMW, around 250 million kilometres of testing and analysis is seen as the required distance to prove and improve the technology. “The mistake some companies have made is to enable that technology straight away. And then, of course, things can go wrong. We will enable the technology as its robustness develops in the next three or four years.”
In 2017, BMW announced testing of adapted 7 Series vehicles in Silicon Valley, in partnership with Intel and Mobileye as part of a plan ‘to bring solutions for highly and fully automated driving into series production by 2021’. Regulators naturally want testing to prove the technology, but some authorities are more open than others to autonomous vehicles being tested on their public roads. Is Robertson satisfied with the support the industry is receiving from municipalities and local or national governments on this technology?
One person’s excitement is another person’s tedious drive. I think the technology will suit different people in different circumstances
“We’re at a very early stage of the regulatory framework. Many governments, such as the UK, Germany, the US and China, are encouraging the pilot application of this type of technology with the very clear statement that there has to be a driver in control when needed. The piloting of the technology is enabled in several countries now. So, that’s useful. And as I said, the ability to reach Level 4 and Level 5, where the driver can do other things, or even get in the back, is something that no regulator has fully accommodated.
“But they’re enabling the pilot because the pilot has, in essence, limited risk. In such a situation, an engineer would be seated behind the wheel working on developing the technology but also on hand to be a driver should the need arise.”
With no-one driving, the question arises as to how people might occupy themselves in their autonomous cars. The notion that people might sit in autonomous cars and work, or read, or watch movies projected onto the car’s windows is countered by the fact that a small but significant number of people suffer some degree of motion sickness.
Level 3 is probably not possible under all circumstances. Our view is that you need to develop the technology through Level 3 to get to Level 4
The success of the fully autonomous car with a fully customisable interior will depend on understanding and overcoming the causes of motion sickness, an area of research that, due to the lack of conclusive findings, is in its relative infancy. How can vehicle manufacturers accommodate motion sickness and other passenger reactions into the type of use that might be envisaged for a fully automated vehicle when the driver is not driving and people are sitting in the back, maybe watching films on their windows instead of looking through those windows?
Robertson has spoken publicly about the strange sensations he experienced when participating in BMW autonomous vehicle research. “We have a very extensive programme and a very realistic simulator, which we’ve been running now for about three years. I sat in it for what was due to be a two to three hour session, but after 15 minutes I was feeling so ill I had to pull the plug on it, literally.”
There are many ambitious suggestions as to how people might view images and movies or surf the web in their autonomous cars, he notes, “but one thing that is probably now very clear is that if you try to view images when the side panorama is moving past you, it doesn’t work too well.” Trains and planes offer something interesting to look at outside, including sight of the horizon; occupants also generally do not sit facing the window, they only face forwards or backwards, and travel with the direction or face in the opposite direction of travel. Furthermore, trains and planes rarely brake suddenly, and their long history as established means of transportation includes deep familiarity with their respective travelling sensations.
“From the point of view of what you can do as a passenger in an autonomous car, the idea of moving the seats so they face each other right to left isn’t going to happen. Moving the seats so you can look at a screen that is on the side window probably isn’t going to happen either. The direction of travel will therefore be the direction in which you can do other things, such as watch a movie or read a book, or use some other device. But the flexibility that many people were thinking about is not possible because motion sickness will prevent it.”
Are we there yet? (No.)
It’s often said that autonomous drive technology is ready, with the mass rollout of autonomous cars being held back merely by the regulators. Robertson, however, disagrees. Asked whether the technology currently available is sufficient to bring autonomous vehicles successfully to market, his reply is short and sweet: “No, it’s not.”
This explains why BMW has said it will take at least another four years to develop the earliest deployable technology. “And for Level 5, even longer.” Is Level 5 really a viable proposition in the foreseeable future? “It’s basically the amount of data and the artificial intelligence to interpret it, and the systems to manage it securely, that require development,” explains Robertson. “And we know from the vision systems of 18 months ago to the vision systems of today, that we’re making massive progress, but we’re still not there. It’s still not enough. Compared to flying a plane across the Atlantic, which has very few variables, driving a car, in most circumstances, has thousands and thousands of variables.”
Underlining the complexity of developing this technology, Robertson ends on a note of caution and reality: “At the end of the day, accidents happen because of human error. We will see many fewer accidents because human error will be avoided. What we don’t have yet is the technology to simulate the capability of the human to make a decision that, in many instances, avoids an accident. And of course, that same decision-making in some instances unfortunately results in an accident.”
This article appeared in the Q3 2017 issue of Automotive Megatrends Magazine. Follow this link to download the full issue