ページの本文へ

Hitachi

企業情報研究開発

2020年1月29日

AIがどのようにモビリティ領域の変革をもたらしているか、Connected Places CatapultのEd Mayo氏とドイツ人工知能研究センター(DFKI) オーグメンティッド・リアリティー研究部のDidier Stricker教授を招き、日立ヨーロッパ社 Social Innovation部門のJim Donaldson、同社 Automotive and IndustryラボのMassimiliano Lenardi、Experience DesignラボのGiulio MoffaとBig DataラボのNick Blakeがそれぞれの立場から現状と今後の展望について語った。

Shifts driving new models of mobility

Nick
I’d like to kick things off by getting your perspective on some of the big political, economic, social, technical, environmental and regulatory shifts (as reflected in the UN’s Sustainable Development Goals, Society 5.0) that are driving new models of mobility that we’re seeing.

Didier
Currently, the majority of cars are driven by one person travelling about 20 or 30 km a day in the city. In the US, they have a special High Occupancy Vehicle (HOV) lane where you can only drive if there are two or more occupants in the car; so you can drive on this pretty empty lane which encourages car sharing journeys into the city. This increases mobility so I think we need to work on this kind of sharing.

Max
Another important shift - especially because Didier works in DFKI where Artificial Intelligence and the importance of data is paramount – is the fact that all forms of transportation, and even pedestrians, are connected and capable of providing a lot of data relating to mobility.

Didier
I completely agree that monitoring movement is very important because then you can run simulations and make predictions using real-time data to optimise the traffic system and mobility options accordingly.

Nick
Ed, your organisation does a lot of work in this area sponsored through the UK Government’s Innovate UK programme, now part of UK Research and Innovation. What do you think is driving this revolution in connected and autonomous mobility and what are the benefits?

Ed
I think it’s several fold – firstly getting people out of their personal vehicles onto public transport or shared transport as much as they can, or at least increase throughputs on the road network, for two benefits: one is reduced emissions and the second is the economic benefits of getting people from A to B more efficiently. It’s very much about helping the environment by reducing vehicles on the road and maximising throughput for the greater good.

Nick
Giulio, I know you’re keen to come in with your experience of the social aspects of these shifts.

Giulio
Yes, in Europe especially the way of working has changed. People have more freedom of movement, often travelling and working overseas, so owning a car with all the associated costs has actually reduced the appeal for younger generations to own a car. But if you move from ownership to something that is shared, there are unfortunately more issues around vandalism and theft.

Connected, autonomous, shared and electric

Nick
We often talk about the four areas of vehicle innovation as being connected, autonomous, shared and electric. Let’s discuss some of these in more detail. Firstly, electrification has raised a number of issues in terms of the infrastructure needed to charge vehicles and minimise the impact on the ageing network infrastructure. What are your perspectives on the adoption of EV? Has it been as rapid as we would expect? Is it a segue to something else? Are EVs going to make it right through to critical mass or are we going to see a segue into other forms of power train?

Jim
From my perspective, I think the initial EV uptake will be driven by fleet rather than domestic. The majority of new car sales go to corporate – something in the region of 60% - and the positive economics associated with EV such as lower fuel costs, lower maintenance and no congestion charging means that there is an increasingly positive business case for transitioning to EV.

Max
Connectivity is already happening, and even low-end vehicles are connected using 4G. The infrastructure is getting more and more connected with sensors along roads, rails and in the ports; connectivity for me is something that is almost achieved in that sense. But as you are probably aware, there are many levels of autonomy for vehicles. Currently, I would say mass-market products are at level 1 and eventually level 2 with some functions but there are 5 levels and only the 5th one corresponds to full autonomy (no steering wheel). It’s a very long way to the 5th level on mass production, especially if we think about complex urban use cases, and there will be a long transition period where several levels of autonomy will be present.
But regarding the introduction of hybrid or fully electric vehicles to the market, it’s not properly taken off firstly because, despite incentives from the government, prices are not yet attractive for the masses, and secondly that people are hesitating to buy a true electric car simply because if they need to travel further, it wouldn’t be easy to recharge.

Jim
As Max said, there are many challenges to overcome before we reach the fully autonomous level 5 but there are some quick wins. For example, one of our fleet customers ran a trial utilising active braking systems which showed a reduction of rear-end collisions of 20% resulting in a positive business case. As vehicles become more sophisticated, we can see a future where we slowly migrate towards a fully autonomous future, but still, deliver value in stages along the way.

Connectivity and autonomous vehicles

Nick
It would be good to get Didier’s perspective on the subject of connectivity because it seems to me there are a number of significant barriers to the adoption of autonomous vehicles; one of those is around how you migrate from a hybrid environment when you’ve got human-driven and autonomous cars on the same roads. Equipping autonomous vehicles to co-exist with human-driven cars is quite a challenge from a technology, trust and safety perspective. The second issue is around connectivity, because until you can optimise the flow of traffic from a central point, dispensing with traffic lights and road signals and so on, you’re not going to get the breakthrough benefits from fully autonomous vehicles. A final thought on this, if we’re going to get completely connected systems, it requires automotive manufacturers to agree on the standards for that interconnectivity. Who’s going to deliver and operate the required citywide or national connectivity? Will it be the manufacturers, or will they need to form consortia to do that? Didier, I guess there are a number of questions here, what’s your perspective based on your work in the area of the connected city?

Didier
I do agree cars will need very good connectivity capabilities. Firstly, to get all the information they need in real-time. It’s not enough to just get information from the sensor of the car itself, you need to know what is in front of the car that is in front of you to anticipate better. We’ll need connectivity just to update the data and the maps so the car will receive real-time data on the general situation to optimise driving trajectories.

I think the breakthrough will come when you have connectivity with all the security aspects and so on, but it’s also the car is becoming a software product itself so you need it to connect and update, that’s the last part I think. So, for the car companies, it is a big shift. With connectivity you need the infrastructure, you need the cloud and at the moment the big software companies are used to it but not the car companies.

Barriers to widespread uptake of EV

Ed
Just going back to Max’s point on the electric vehicles and generational shifts and so forth, I think what’s stopping people buying electric vehicles at the moment is probably range anxiety. As Max points out, most car journeys are really short, so that’s almost a misconception by the general public. But I think the other thing that holds people back is the charging infrastructure – can it be relied upon and do we know where they all are? Secondary to that are probably price and availability, and the very small second-hand market.

Jim
The transition to EV is bringing together two previously unconnected ecosystems: mobility and energy. This is the reason that we are leading the Optimise Prime project, which is our flagship EV programme where we are exploring the challenges for both the electricity system and fleet operators. In terms of infrastructure, it’s not yet clear what that EV charging infrastructure should look like.

In a world of car ownership, it’s likely that the charging infrastructure would be located at homes and offices so we are optimising to maximise the customer experience. In a world of shared mobility and robot taxis, we could move to centralised charging hubs at sites that are optimal from an energy system perspective. We really need to think at system level to determine the optimal solution for society rather than individual organisations or stakeholders.

Multi-modal transport and infrastructure

Nick
Let’s switch track a little from the “CASE” discussion to a wider conversation. In this scenario, connected and autonomous vehicles become an integral part of the solution to the challenges of urban and rural transportation. One of the issues that we note in the UK, and I think it exists elsewhere, is that as cities morph into conurbations, the traditional hub and spoke transportation systems start to breakdown as they were designed to support the rising need for point-to-point travel across the conurbation. This is a real barrier to growth and productivity in areas of the UK such as the Manchester, Leeds, Sheffield area. So, it seems to me that connective autonomous shared electric cars are only part of the solution to the emerging urban mobility problem where aging mass transportation systems are starting to break under pressure. Guilio?

Giulio
Well as I was saying before, if we see autonomous driving as a part of the last mile journey within the mobility-as-a-service scenario, and we have long distances covered by public transportation and rail for example, so we might see those points as being the real issue. For example, the long-distance cover with batteries is one of the barriers. Of course, this requires an infrastructure that can cover the whole area including the remote rural part, not just an edge of a city, with efficiency. And definitely, there may be some potential in new sharing services to help make this transformation faster.

Nick
So as journey requirements change, cutting across into traditional city "centres of gravity", these journeys increasingly involve complex mobility mode and route choices and this presents a range of challenges such as how to get from A to B quickly and cheaply, avoiding traffic bottlenecks within a single, simple integrated service and payment framework. That sounds like an opportunity!

Giulio
Absolutely! So eventually there will be a change in the concept of tickets, perhaps paying a monthly fee for travel like you do for things like Netflix; you have access to the service, so you don’t need to own it. You will always need transportation and, of course, there will always be the requirement of owning a vehicle but owning a vehicle is not the requirement of mobility.

Jim
I agree. Ultimately, we need to focus on customers’ needs, and look at how best to fulfil those, rather than being wedded to existing technologies and business models.

Global innovation

Nick
OK, so if we look at three big regions of the world – North America, Europe and Asia-Pacific, particularly China – where do you think the breakthrough innovation is coming from? Are the innovation “hot-spots” the same as the adoption “hot-spots”? For example, innovation may be coming from Europe or North America but the market for mass adoption might be in new cities in China?

Ed
In those three regions, they’ve got different challenges and advantages over the others. If we take North America for example, they have a large landmass so in some ways they’ve got more scope for innovation and application of innovation because they have more geographical room to play with. Take, for example, air taxis in North America, you’ve got more room to take off and land and that’s probably a concept that will be far more difficult to implement in Europe. I don’t know China, but I would assume they have a similar sort of issue with very large populations and conurbations. So I think some of the more difficult requirements will probably come from Europe purely down to its geographical nature and the existing transport infrastructure, but in terms of getting to market first, I think probably North America stands a better chance because they’ve just got more space to play with.

Global adoption

Nick
So just staying on that subject a little bit, we’ve spoken about the innovation, but the adoption is a different thing. Now, adoption may vary according to regional social predispositions and regulatory environments. If, for example, China has a looser regulatory environment for these new forms of mobility, does it mean that early adoption is more likely there? And if that is the case, will the natural teething problems be ironed out there first, leaving ageing European travel and mobility systems behind?

Giulio
The biggest example of this is Uber; over the last few years they have encountered tons of regulatory problems in Europe, so every country in Europe applies Uber differently and this is one of the big areas where, even if it is a global service by itself so you can use your account, you will receive a different service in each country. Regulations present challenges so I think it’s back and forth with regulation slowing adoption of innovation.

Ed
I think there tend to be two drags on technology; one tends to be regulatory and the second tends to be infrastructure, and unfortunately both are quite slow-moving beasts.

Commercial mobility

Nick
We have spoken a lot about passenger mobility and it seems to me that a similar and very significant revolution is also occurring in the mobility of goods and the impact that will have on our supply chains. Max, have you got any perspective on how this type of innovation is going to affect traditional freight and supply chain models?

Max
We’re already working on services for some of our fleet manager customers where we provide algorithms and ‘technologies as a service’ in order to optimise the different aspects of goods transportation, not only the delivery chain and the supply chain but also fleet operation and fleet cost of optimisation in some cases. They are combining data coming from the fleet itself coupled with data from related infrastructures, such as smart charging infrastructure or traffic management systems. So, putting all this data together might have a significant impact on the mobility of goods.

Jim
That’s right, we’re currently working with fleet and transport customers to optimise their operations in terms of allocation and management of vehicles, maintenance scheduling, spare parts management and predictive maintenance plus EV charging optimisation and planning. By taking this ‘whole system’ approach, we can create additional value whilst providing a simpler solution for our customers in relation to road, rail and shipping transportation.

Drone technology

Nick
Everybody’s talking about what’s going to happen with drone technology and whether its mass adoption will be allowed by governments and regulators – there have been some high profile incidents in Europe recently and elsewhere – are we ever going to see automated personal mobility get off the ground and into the air?

Didier
Yes sure, you already have autonomous vehicles being used in massive logistics areas like BASF’s storage facility. Inside, they automate the offloading and transportation of chemical components via tracks that are remotely controlled via a control room. So in the area of logistics it will come first and then, of course, it will expand into other applications.

Concerning the drones, I think it makes a lot of sense to use drones in emergency cases, like first responders, and for medical and hospital applications where it would be much easier to transfer casualties to hospital via drone than a helicopter. And maybe also taxis for travelling from A to B in the city.

Innovations for a mobile future

Nick
Are there any other kinds of innovation out there that are on people’s radars that could form the next wave of innovation beyond the connective, autonomous and electric vehicles? For example, the concept of autonomous pods that serve the first and last mile but plug into or go inside larger autonomous forms of transportation?

Ed
I think in order for that kind of technology to work, it requires a lot of data and understanding people’s journeys, and that underpins air taxis and the train-like concept you just described Nick, where you have a number of pods joined together and then they splinter off as they reach their destination points en-route. In terms of understanding people’s movements, I think that’s key to a lot of this new technology and whether it will ever be adopted or taken up; companies like Addison Lee and Uber will know an awful lot about people’s movements across the cityscape and they must have a ton of very valuable data to inform the understanding of the future of mobility, challenges and how they might be met. Overlay tracked electronic device data onto modes of transport, then you can start to deduce how people are moving about, how long they stay at places and what they’re doing there. So, I think there is a massive amount of data mining and intelligence gathering that could help inform future transport modes, and how they might be best optimised and used.

Giulio
Regarding drones, I think years ago there was a form of concept by DHL about logistics to go to islands, say, where there’s no existing infrastructure or maybe limited ferry crossings. Or there may be some niche technology, like the Hyperloop for example, where you’d have to build infrastructure from scratch, and you need really to go fast from A to B.

Nick
Well, I think the reality is that certain parts of the world have particular spatial and social demographics that lend themselves more to the Hyperloop concept than others. It’s difficult to see how you would make that work in Europe but if you look at North America which is criss-crossed with interstate highways, you could potentially run the Hyperloop infrastructure alongside or above the road. The Hyperloop idea fits quite nicely with the pod concept that I was talking about – autonomous pods could pick passengers up from homes or offices and load them automatically into the Hyperloop which then transports the pod at high speed to the drop-off point at the next city for the last mile.

Role of AI in autonomous vehicles

Nick
I don’t think we can have this conversation without having a discussion on the role of Artificial Intelligence, which is of course at the centre of this innovation. There has been a lot of debate about safety, trust, adoption and accountability. Max you have alluded to this issue of trust and also that the OEMs tend to be quite engineering-led. Will they ever trust an artificial intelligence-centred autonomous vehicle? In our R&D projects, we are using artificial intelligence for motion control, but we’re also now heavily into research using artificial intelligence for advanced perception. Do you believe that artificial intelligence in its current state is sufficiently advanced and mature to support this level of autonomy in vehicles, or whether it needs further maturity before we see a real adoption and acceptance of autonomous driving?

Didier
In the last year, we’ve had a kind of explosion of enthusiasm about AI and we’ve made a lot of progress in the sensor analysis, data analysis and image analysis, but what is also interesting is that that this area was only driven by one technology. But we will need quite a long period of time for validation and crystallisation, so we have, let’s say, demonstrated the first proof of concept. How to road test and make this technology very robust over a long period of time is still an open question.

Max
I think that advancing technology is allowing us to emulate systems more accurately and there are more and more start-ups appearing, modelling from small sets of data. The quality of their models has to be proven totally but, in my opinion, we should find ways to shorten and accelerate this learning phase much more in collaboration with research teams and private companies.

Data and digital trust

Nick
Just moving on to some of the other barriers, there’s a big elephant in the room and it’s around trust. We’ve spoken about trust in the use of AI but there are two other aspects of digital trust that appear to me to be significant barriers to the widespread adoption of connected and automated vehicle technology. One is around data protection, in particular Europe with the GDPR. The second one relates to protection against the threat of cyber attacks. This threat exists for all IoT-enabled automation, but of course, it is a particularly sensitive issue when it comes to automated vehicles. So, a couple of big issues that I would appreciate having your perspectives on, maybe starting with Ed because you’re facing these issues in your innovation programmes here and now.

Ed
Yes, I think confidence and trust are key: do you trust the system, and do you have the confidence in its capabilities and ability? I think they’re two massive factors because they lead to general acceptance by the public. I think if people can be shown that there are benefits to be had by sharing this data, then that’s one way to win people over. Or you might even be able to monetise and incentivise to encourage people.

In terms of cyber security, it’s such a massive area and has such potential for malicious forces to cause trouble as we become more and more connected. It’s very important for industry and business to show how they’re protecting the systems which will improve people’s trust and confidence in them, but it’s a difficult sell and it’s a difficult thing to prove.

Nick
That’s a useful perspective, thank you, Ed. Any other thoughts on this very complicated issue?

Jim
I think we’re still early in the journey of understanding the potential impacts on society from mass accumulation of data by private organisations and over the last couple of years there have been numerous situations of personal data being misused, stored against permission, cyber theft, hacking and monetary loss for individuals and, of course, the Cambridge Analytica scandal. When we are talking about connected vehicles, we should be clear about why data is being collected and be able to demonstrate clear value to society. If we can’t, it shouldn’t be collected.

Social innovation vs economics

Nick
This conversation started with a perspective that one of the key drivers for this innovation revolution was around reducing carbon emissions. We’re still in a business environment where companies are valued based on their short-term profitability. Most corporates now have a significant corporate social responsibility (CSR) dimension to their strategy, but it still looks a little bit like an add-on – do we need to move towards balancing our values as corporations between the economic, social and environmental dimensions if we are to succeed to in driving this innovation through in a balanced and sustainable way?

Giulio
It’s a big question of course, and the short answer is it also goes hand-in-hand with trust. So, if the impact of this new technology will be good for the environment in the long term, it will be much more beneficial to make the necessary investment.

Max
We have to work towards social innovation business in the sense that we have to think about solutions that not only bring in profit but also tackle social issues, like pollution. If you think about the autonomy aspect of transportation, most people in the industry say that this will automatically bring about a reduction in pollution through better traffic management and better use of different type of engines, but I’m talking about the phase where there will be a mixture of steel combustion engines and electrified ones. If we don’t meet passenger expectations, then these new vehicles will not sell (despite the environmental benefits) so business will not be happy. So, we have to look from a technology point of view to satisfy end users requirements, like for example, comfort in the autonomous vehicle, I think.

Didier
I think that’s why sometimes you need regulation and political engagement in combination. But you have to make sure that the political debate is as fast as the technology revolution or the need for new solutions; that’s my concern and in many projects now, at least in Germany, we try to involve a partner from social science in the project so we are having a discussion about culture and social science together with innovation.

Nick
So in wrapping up the discussion, are there any final thoughts that we’d like to share?

Jim
I think we’ll see changes in how cars are used, shared and fuelled. Equally, we’ll start to see a rise in charging stations which are co-located with other renewable energy assets – the increased demand by the vehicles will be offset by additional, low-carbon generation. One area where it is less clear is for HGVs, and hydrogen seems to be a viable solution here albeit less efficient than pure electric over the full cycle.

Nick
It seems to me that in order to achieve mass adoption of electronic vehicles, we need to make the process of charging these vehicles as easy as possible; and that means putting charging facilities in the right locations; as well as having intelligent services to optimise the charging, storage and depletion of batteries. We know that the mass adoption of EVs is going to have a significant impact on our energy generation, transmission, distribution and balancing systems, that are themselves being significantly impacted by increased volatility of supply as a result of the rise in intermittent sources of renewable energy, so it is a big, big issue.

Of course, it is possible that the market for EVs will itself be overtaken by new forms of energy storage, for example advances in hydrogen fuel technology that appears to be a focus for innovation in China. I guess we will have to wait and see. But I guess whichever direction of travel we take, there is a movement towards carbon-neutral power trains and this revolution will have significant impact on the global car industry as well as on our society.

Many thanks to you all for what I think has been both a broad ranging and insightful discussion.

In the next article in this series, Hitachi Researchers will be joined by Professor Yoshua Bengio, founder of the Montreal Institute for Learning Algorithms (MILA). The discussion will focus the societal impact of artificial intelligence. Learn how Hitachi is working with MILA to harness the power of AI for the benefit of society.

This article was prepared by the European R&D Centre, Hitachi Europe Ltd.

プロフィール

※ 所属、役職は公開当時のものです。

Didier Stricker, Ph.D.

Scientific Director and Head of the Augmented Reality Research Dept.
Deutsches Forschungszentrum für Künstliche Intelligenz (DFKI)
Professor of Computer Science
Technische Universität Kaiserslautern

Prof. Dr Stricker studied electrical engineering at the Technical University of Grenoble as well as at the TH Karlsruhe and graduated at both universities. In 2002 he received his doctorate at the Technical University of Darmstadt for: "Computer Vision-based Calibration and Tracking Methods for Augmented Reality Applications".

He works as an expert for various European and national research organizations. His scientific focus is virtual and augmented reality, computer vision, human-computer interaction and user recognition.

Ed Mayo

Programme Manager
Connected Places Catapult
Hitachi Vantara

Ed is a Programme Manager with complementary solution architect skills in defence, civil aerospace markets and automotive, accumulated over 30 years. He started his career as an Aerodynamicist (working on Harrier) moving through operational research, simulation and training and, since 2000, into the emerging field of autonomous systems including drones and CAVs (Connected Autonomous Vehicles).

His career has been spent in the scientific, engineering and technical field both innovating and delivering to end users. The last few years have been spent helping deliver CAV demonstration programmes working with industry, SMEs and academia to help position the UK as one of the frontrunners in the CAV R&D space.

Jim Donaldson

Director of Innovation,
Hitachi Social Innovation EMEA
Hitachi Europe Ltd.

Jim has over 15 years’ experience delivering IoT, AI and Big Data-based solutions in Security, Software and Energy systems and holds 15 patents. Prior to joining Hitachi, Jim spent 10 years as CTO of an Oxford University spin-out company developing cloud-based AI driven energy management systems for domestic households, with a focus specifically on reducing energy and carbon consumption in the home.

As Technology Lead for a number of Hitachi’s Social Innovation projects, Jim brings together Hitachi’s technologies in AI and IoT to deliver solutions in Energy, Smart Transport, Zero Carbon and Smart Cities. Jim is a chartered engineer and holds a 1st Class MEng degree in Engineering Sciences from Oxford University.

Massimiliano (Max) Lenardi, Ph.D.

R&D Laboratory Director and Senior Research Engineer
Automotive and Industry Laboratory (A&IL)
European R&D Centre
Hitachi Europe Ltd.

Massimiliano (Max) manages the "Automotive and Industry Laboratory" (A&IL) across its two main sites in Sophia Antipolis, France and Munich, Germany.

In the Automotive domain, Max’s team innovates with ICT and smart automotive components for Dynamic Control of vehicles, in particular Connected Autonomous Vehicles (CAV), and with smart engine combustion technologies.

In the Industrial production domain, his team innovates smart manufacturing processes with ICT and Industry 4.0 concepts, also cooperating with DFKI on Artificial Intelligence (AI) applied to Human Activity Recognition.

Max has been a Coordinator of the EU FP7 project "eCo-FEV", and his teams successfully contributed to the Autonet2030 and DRIVE C2X European projects. The Hitachi R&D labs are currently an active partner in the DENSE and HumanDrive publicly-funded projects.

Giulio Moffa

Senior Designer
Experience Design Lab
European R&D Centre
Hitachi Europe Ltd.

Giulio joined the Hitachi Design Centre in Milan in 2007 after having worked previously with Piaggio, Samsung Electronics and various design consultancies in Italy. Giulio has built up extensive experience in both product and service design projects across different areas of the Hitachi business after moving to Hitachi Europe in 2010. Currently, Giulio is in charge of Rail and Mobility projects in Europe and is using ‘user centred’ design methods to validate ideas from multiple perspectives.

Giulio has a BA and MA in Design from ISIA Roma and Domus Academy.

Nick Blake

Chief Innovation Strategist and Head of European Big Data Lab
European R&D Centre
Hitachi Europe Ltd.

Nick is responsible for driving sustainable and scalable innovation in IoT, Artificial Intelligence and Automation through collaboration with customers, corporate ventures and academic institutions working across a range of sectors including Transport, Automotive, Manufacturing, Healthcare, Energy and Public Safety. His current work covers autonomous driving, advanced video intelligence, smart spaces and data protection.

Previously, Nick built a market-leading big data and analytics business at Capgemini, focusing on delivering advanced analytics across the areas of consumer, asset, assurance and fraud. His experience covers strategy and transformation across a wide range of sectors including central government, travel and transport, automotive and consumer goods.