TU-Automotive Europe 2017: Day 2

Continued from Day 1

It is now almost common wisdom in the wider automotive ecosystem – and it was stated on both days of the TU-Automotive Europe conference in Munich – that the holy grail of the disruption currently transforming the auto industry is the marriage of mobility-as-a-service and autonomous driving.

As Mark Thomas, vice president of Marketing at RideCell put it: “The future is autonomous ride-haling.” Georg Obermayer, operations and logistics manager for Germany at Uber, agreed. “We think autonomous vehicles will amplify [the percentage of trips requested via an app]. These technologies go hand in hand. It doesn’t make sense economically to own those vehicles and just use them for yourself. This is why we think autonomous will be the next big step in ride-haling.”

He added that Uber currently has autonomous cars live in four US cities and are partners with Volvo and Ford. The company has also just announced that Daimler will put their self-driving E-class vehicles on the Uber platform. According to Thomas, overall about 1% of all passenger miles driven today are in new mobility means, such as car-sharing and ride-sharing. “So we are just at the very beginning,” he says. “By 2030 it’s expected that about a third of all passenger miles will be via new mobility services. Roughly 10% of those will be driven. The real growth engine here comes from autonomous mobility services. That’s because when you take the cost of the driver out of the equation, it brings the cost down to almost below half of the cost of individual vehicle ownership.”

The aim behind mobility-as-a-service is, clearly, to reduce private car ownership and, thereby, reduce the number of cars on city streets. This will reduce pollution, traffic congestion and, just as important, reduce the amount of space taken up by cars sitting idle and taking up space that could be used more productively. “About 95% of the time cars just sit there, idle,” Obermayer said. “This leads to 20% of the space in cities [being] used for parking. Not schools. Not housing. Not parks but parking lots and parking spots.”

Mobility-as-a-service also brings many benefits to users, including, as Thomas pointed out, ease of payment. “The new mobility services are actually roll-ups of multiple industries,” he explained. “Instead of having to pay for insurance yourself, paying to have your car washed, to have it maintained, to pay for repairs – all that just gets taken care of when you get mobility-as-a-service. The service provider provides all that for you.”

In addition, he said fears that the decline and fall of car ownership will have a major negative impact on car sales are simply misplaced. “Automotive sales aren’t going to be diminishing much but who’s going to be buying the cars is going to change the nature of the automotive industry. Somewhere between 2025 and 2030, in North American urban areas, more than half of the cars sold will be to these new mobility services, and not to private individuals.”

If true, this is good news for carmakers and municipal traffic management divisions struggling with overloaded and inefficient transport systems. “When ride-haling took off,” Obermayer said. “We realised that the status quo of the transportation system is inefficient, insufficient and unequal. So the question is, is there an alternative to a world that looks like a parking lot and moves like a traffic jam? We think that if people start to share cars and use public transportation, we’re on a pretty good path.”

Fortunately, people going from a point A to a point B don’t really care how they get there. They simply want to get there comfortably, safely and as quickly as possible. Obermayer described a use case which illustrates the inherent flexibility of transportation users.

“Riders love combining [mobility] options,” he said. “London recently launched the Night Tube. As soon as it was launched in London, in the city centre our business went really down, from the first day this Night Tube went live because Londoners finally had an alternative – maybe even a better alternative than taking an Uber – in the City. But we also saw that in the outskirts our business was booming because people combined things. They take the subway to the station nearest to their home but, for the last one or two miles, they get an Uber and ride home safely.”

Mobility-as-a-service is also inexpensive – and it can be made even cheaper through on-demand carpooling. “If you make one trip out of two or even three, there are a lot of advantages,” Obermayer said. “For the rider, you get a cheaper ride. For the driver, you have more paying minutes per hour, you have higher revenues, better utilisation. For cities, you transport more people with fewer cars.”

He said Uber’s carpooling service has been an unexpected success. “In San Francisco, every second trip is pooled now.” He showed slides that illustrated how traffic volume in the city’s busy Market Street area declined after uberPOOL was launched. “On a global scale, that leads to fewer miles driven, less fuel consumption, less carbon emission,” Obermayer said.

Eventually, the autonomous car will make mobility-as-a-service as seamless an experience as walking out of your front door and hopping into your own vehicle, Thomas believes. “When you think about autonomous mobility-as-a-service, the difference between car-sharing and ride-sharing goes away,” he noted. “Once it’s autonomous, there’s no need to walk to the vehicle anymore; the vehicle just comes to you.”

How far are we from the advent of the truly autonomous car? Predictions vary but most seem to think 2030 is a likely tipping point. In the meantime, as Matthieu Noel, Paris manager at Ptolemus, said: “Level 3 autonomy vehicles have been recently launched on the market with Audi introducing its new A8 model in Germany. We expect the first Level 4 vehicle to be launched on the road by 2021. At Ptolemus, we expect 60M vehicles on the road will be equipped with Level 3 or 4 technologies by 2030.”

The new technology must be tested and retested to ensure its reliability and safety. Testing is now widespread in Europe and North America but, as Noel pointed out, while moderating a panel discussion titled Autonomous Testing: An Action Plan for Europe, “Today, testing activity is more advanced in North America than in Europe. In Europe we mainly see testing from point A to point B on dedicated roads or in closed environments, while in the US we see more testing on open roads.”

The reason for the difference is regulation. According to Harri Santamala, CEO of Sensible 4, “A big issue is a lack of common regulations in Europe. Finland has a liberal approach, as do Estonia, Greece, the Netherlands. There are many places where you can do things quite openly and freely, which helps in speeding up startups and new initiatives. The problem is that country by country things vary, and requirements vary, and I think the traditional approach of requiring type-approved vehicles, which on Level 4 doesn’t exist, limits innovation.”

The reason for the differing degrees of regulation in Europe is political, he said.  “I think the countries that have strong local OEMs have to do what the OEMs see as the best approach to make a lot of sales and the countries that don’t have national car brands, like Finland, can be quite open.”

While Santamala agrees that early testing of autonomous systems should be carried out in closed areas, he notes that “there are real-life things which are difficult to test in those environments and at some point you need to go open road. As a startup we don’t have the resources of a OEM, so it means that we go to the open road earlier.”

The variance in regulations is not the only country-by-country difference bedevilling the testing of autonomous cars in Europe. There is also the variance in broadband network coverage. According to Stephan Joest, account director for device and application verification services in Europe at Ericsson, “We know that we have only 3G or even only 2G and, maybe, even no coverage at all. We know that a lot of OEMs and Tier 1s now have to drive to certain places in Europe to find certain conditions – for example, for a handover scenario, going from one technology to another one, or even in a roaming situation going from one country to another.”

That is obviously inefficient but this coverage variant can also play havoc with the testing itself. “One of the most critical elements we see today is that you find out that the system starts to hang,” he said. “For example, when you have communication and you come from a German 4G environment to a Belgian 3G environment, you might lose all communication.”

This needs, first, to be detected and then the situation must be attended to, all of which takes up valuable time. “It can be that the system has not had any connection for 20, 30 or 40 seconds,” Joest said. “In the worst case, you need to do a reset of your TCU. That might take another 20 seconds. That means you are out of communication for one minute. The question is, can your brand afford that? Can your system, your technology, your architecture, your application, your business intelligence afford that situation?”

His solution is to recognize the problem and then retest again and again. “You drive around with the car, and maybe you have 3G here and LTE there, or maybe 2G-2G only, and test over and over again with real vehicles with a real antenna in the car, so that you can reduce the number of problems to a minimum, to come to the desired position of not only zero fatalities but also zero incidents in your own system architecture.”

Santamala’s firm is currently testing Level 4 technology and they are taking strict precautions to make sure there is always ideal connectivity. “Our approach is there are certain situations that are really difficult to go with the technology.” he explained. “If we are talking about a shared mobility and public transportation business model, we assume it’s like an elevator business model: you take the driver out but you still have somebody in control. For that reason, you need to have connectivity from at least two different network suppliers at all times and if we are talking about driving from digital maps, we don’t store the map in the vehicle. We constantly download and upload the map data in the vehicle.”

It is not only autonomous technology that is being tested; user acceptance of the systems is also being investigated in Europe. Benedikt van dem Boom, project officer at FIA, was on hand to describe a specific project currently being designed for deployment in early 2018.

The project, titled Autopilot, is EU-funded and has sites in Finland, France, Italy, the Netherlands, South Korea and Spain. For the project, FIA has 44 partners from the auto industry, research, insurance associations and public authorities.

The project wants to determine “whether use of the IoT can enhance or accelerate user acceptance of automated driving,” van dem Boom explained. “Put concretely, how can the Iot add some value to automated driving? The IoT ecosystem includes “individuals with smartphones, smart watches, even with smart glasses, and an additional number of sensors that we add [that] can all provide traffic-relevant data into the cloud and can be used for autonomous driving,” he said. “We attempt to virtually recreate these objects in the IoT platform that can be interoperable across different service providers.”

He named two pilot sites used in the project. One of them is the French city of Versailles. “We take one-seater Renault Twizy cars that can be driven by tourists manually across the city of Versailles until they reach the gardens of the castle,” he said. “There they are encouraged to switch on Level 4 driving functions so that the car drives completely autonomously through the gardens of Versailles. To avoid accidents and other unpleasant encounters, the car will pick up signals from other cars and tourists in the gardens – for example, cyclists that wear smart watches.”

Another pilot site is Livorno, Italy, where Level 3 highway driving will be done. “Using standard IoT protocols, we collect data from the road – for example, about puddles, roadworks, about vehicles that might be blocking the way,” van dem Boom said. “This data feeds into the national traffic control centre and is processed until it shows a warning indication to the driver in the car.” Several hypotheses will be tested in the project, he noted. “One hypothesis we are very keen on testing is whether the visualisation of IoT infrastructure increases the acceptance. Put bluntly, if users see which data goes where, are they more accepting of an IoT-driven automated driving function?”

Another hypothesis the project is examining is whether the cost of the IoT infrastructure is too expensive vis-à-vis the value IoT brings to autonomous driving. “That’s a very negative and pessimistic hypothesis,” van dem Boom said. “We hope to be proven wrong. It’s not clear whether we need the IoT for autonomous driving in the first place. We want to find out if it can have an added benefit on the economic level and on the usability level. This is at the heart of our evaluation strategy.”

Leave a comment

Your email address will not be published. Required fields are marked *