Intelligent Transport Systems (ITS)

As we ride the wave of digitisation and become increasingly dependent on technology, security becomes more and more important. Recent high-profile security breaches have highlighted our vulnerability and as a result a lot of work addressing security has been funded. However, is this work focused on the right areas and threats? From a technical point of view, high levels of security can be achieved, however, the real issue is more a matter of adoption of security solutions by CPS/IoT actors. There a number of different challenges that need to be addressed:

• Development costs and business model – electronics & software-based systems are expected to be very cheap (cost by function is decreasing rapidly for systems). However, the security solution usually needs to be tailored to the platform/function/system, which raises the development and validation costs without increasing the customer added value.
• Limited performance for CPS and IoT – e.g. in terms of bandwidth, computing performance, energy requirements available that limits the security-related counter measures that can be deployed.
• Usability and User acceptance – increasing the security (e.g., authentication, confidentiality, privacy) usually decreases usability and has an impact on user acceptance. Customer training and awareness is needed which is, firstly not always possible, and may decrease the attractiveness of the system in comparison to other non-secure solutions.
• Competences – The design and deployment of security counter measures requires a combination of competences with respect to security engineering on one side, and the product on the other side. There is a lack of qualified engineers who can address these issues.

So, in addition to performing research on security, there is also a need to invest in and train security engineers to think in terms of business, opportunities and ROI. There is also a need to educate high-level management of traditional industrial sectors to make them more aware about risks and opportunities related to security specifically, and ICT in general. Until this happens we cannot be confident that the products we are buying are truly secure, but are we prepared to pay the price for good security?

It is clear that there is a need for regulation with respect to safety, and an update to the laws regarding liability. Although mandatory standards exist, e.g., ISO 26262 for safety in the automotive domain, IEC61508 for industrial electronics, and other more domain specific standards to maintain acceptance on the market, there is currently no public authority third party, such as found in the aerospace industry, to oversee the release of a newly developed product. This responsibility is kept at the OEM level.

Safety regulation is needed, but this needs firstly to recognise that removing human error will on balance save lives. The aim is not to have zero accidents. However, a challenge is that humans are about three orders of magnitude less tolerant of deaths caused by external factors (e.g. train crashes) than by factors they perceive themselves to have control over (e.g. car crashes). As a consequence, autonomous systems need to be about 1000 times safer than the manual ones they replace to be accepted. There is also no point making the owners and operators of autonomous systems legally responsible for systems that they have no chance of understanding or managing. The legal responsibility needs to be shifted from individual vehicle owners and “operators” to the designer and infrastructure system operators.

There is also no point in having autonomous systems that require a human operator to take over in a time-critical emergency, because the human operator will not have sufficient situational awareness to do so. This is shown by numerous aviation studies. It takes a human operator a significant time to re-orientate, so an autonomous vehicle will need to automatically bring itself into a safe and stable state before handing over to a human. Fundamentally, fully autonomous operation is needed to cover the situations where autonomous cars are particularly beneficial, e.g. when the human is drunk, asleep, unfit to drive (too old, blind, etc.).

Looking to the future the expectation is that the insurance industry will shift to insuring the manufacturers of autonomous systems rather than car owners. If companies decide to self-insure, there may be a need for legislation and government to ensure that a company’s maximum potential liability does not exceed their assets. If they do not have sufficient assets, then insurance should be compulsory.

So what is needed? Firstly, there are many restrictions across Europe on autonomous driving and here the European Commission has a role to play in providing harmonisation. This requires regulation, clarification of the responsibility of the OEMs, and support to educate the public, e.g. via public demonstration. Already a number of significant actions are already underway, but there is still a long road ahead before we see fully autonomous vehicles on our streets.

The capability of computer programs is now reaching the same level of performance as humans in areas such as image processing and in other tasks such as calculus or memory, computers have already exceeded human capabilities. There is a general assumption that deep learning could bring important progress in other tasks that deal with unstructured data sets (e.g. data analytics or “big data”). Going beyond this AI has the capability to improve control functions and two aspects need to be considered more carefully:

• Use of AI for safety-critical control strategies – The formal or even systematic verification of an AI algorithm is nearly impossible due to the strong impact of the learning set on the performance and behaviour of the algorithm. The usage of AI for safety-critical systems (e.g., object recognition for autonomous driving) needs to be assessed very carefully to ensure that a proof can be issued that no critical issue might occur.
• Usage of AI for private data – similarly, the behaviour of the algorithm is very difficult to predict, and consequently it is very difficult to enforce that a given regulation can be respected in any situation.

In general, it is very difficult to prove that an AI algorithm will behave within a given regulation, technical, ethical or legal constraint. Additional technically imposed limitations might thus be required to ensure this. These would also contribute to supporting customer trust and acceptance.

However, there are some specific challenging scenarios that need to be considered. For instance, imagine an intelligent building management system that can understand how many people are in different sectors of the building. A fire breaks out and is spreading more rapidly than the rate at which people are evacuating. Does the intelligent system close off a part of the building which is on fire, preventing some people from escaping, but overall allowing a greater number to escape and prevent further damage to the building? Or should it always have the rule to ensure that in the event of a fire all escape routes are left available, no matter what the damage is to human life or to the cost of the building. These may lead to some logic statements that may either:

• Minimise the fatalities and casualties with a scoring system, weighting fatalities against different injury types.
• Minimise cost impact with cost damage to the building considered as well as values given on life and injuries.
• The system is programmed to never take positive action which may endanger a life, even if this potentially results in a higher number of injuries and fatalities.

So regulation for AI is needed such that the logic statements used are subject to tests to see how AI systems respond in a real-world environment. Of course this may raise such challenging ethical questions which will require a lot of debate and supporting regulation to make it clear what decisions are acceptable to society.

In the last blog the coming GDPR regulation was highlighted and the implications that this will have for all of us. On a positive note this is good for consumers. The challenge though is for companies who need to meet this new regulation. Any company that stores or processes personal information about EU citizens within EU states must comply with the GDPR if they have:

• A presence in an EU country, even if they do not have a business presence within the EU
• No presence in the EU, but it processes personal data of European residents
• More than 250 employees
• Fewer than 250 employees but its data-processing impacts the rights and freedoms of data subjects, is not occasional, or includes certain types of sensitive personal data

Or in other words just about all companies. So how does a company comply? A problem is that the GDPR takes a wide view of what constitutes personal identification information so the same level of protection for an individual’s IP address or cookie data will be required as for sensitive data such as name, address and Social Security number. There is also a problem of interpretation. Companies must provide a “reasonable” level of protection for personal data, although what constitutes “reasonable” is not defined. A consequence of this is that there are likely to be quite a few fines for data breaches and non-compliance as GDPR is introduced.

The GDPR defines three key company roles that are responsible for ensuring compliance: data controller, data processor and the data protection officer (DPO). The data controller defines how personal data is processed and the purposes for which it is processed. The controller is also responsible for making sure that outside contractors comply. Data processors may be the internal staff that maintain and process personal data records or any outsourcing firm that performs all or part of these activities. Notably it is the data processors who are liable for breaches or non-compliance. Thus if your cloud provider is fined you may well be fined as well so choose carefully. A DPO needs to be designated to oversee data security strategy and GDPR compliance. Companies are required to have a DPO if they process or store large amounts of EU citizen data, process or store special personal data, regularly monitor data subjects, or are a public authority.

So what if your company is non-compliant? The GDPR allows for penalties of up to €20 million or 4 percent of global annual turnover, whichever is higher. Some predictions are that around half companies will not be compliant when GDPR comes into force and around $6 billion in fines and penalties will be collected in the first year. How these will be assessed, e.g. what is a major breach that could cause damage and what is a minor breach, will need to be decided upon. Here GDPR places a requirement on companies to perform impact assessments to mitigate the risk of breaches by identifying vulnerabilities and how to address them.

The good news for consumers though is that we will get a lot more information about data breaches. A key requirement brought in by GDPR is that companies must report data breaches to supervisory authorities and individuals affected by a breach within 72 hours of when the breach is detected. Thus, although the introduction of GDPR is likely to be costly and painful for companies, we will know a lot more about how safe our data is in the future and also which companies we can trust with our data.

By May 25, 2018 companies that collect data on citizens across all 28 EU member states will need to comply with strict new rules protecting customer data with the introduction of the General Data Protection Regulation. This dictates that it will not be allowed by law to collect data on:

• Basic identity information such as name, address and ID numbers
• Web data such as location, IP address, cookie data and RFID tags
• Health and genetic data
• Biometric data
• Racial or ethnic data
• Political opinions
• Sexual orientation

This is good for consumers as there are clear rules with respect to their data and it is also good for companies as they only need to comply with a single standard within Europe. However, the requirements to meet and administer the standard will require most companies to invest heavily. The large US companies that deal with data are expecting to have to invest significant amounts to meet the new standard. According to the PwC survey, 68 percent of US-based companies expect to spend between $1 million to $10 million to meet GDPR requirements. Another 9 percent expect to spend more than $10 million leading to some complaints that it will put them at a competitive disadvantage with European Companies. As European companies need to abide by the same rules it is not entirely clear where this disadvantage comes from except that the GDPR regulates the exportation of personal data outside of the EU and many US companies have data centres and support staff in the US.

So what does this mean for companies? The GDPR requirements will force companies to change the way they process, store, and protect customers’ personal data. For example, companies will be allowed to store and process personal data only when the individual consents and for “no longer than is necessary for the purposes for which the personal data are processed.” Personal data must also be portable from one company to another, and companies must erase personal data upon request enshrining the concept of the “right to be forgotten”.

So the good news is that from May 25 2018 we will get a bit more privacy - the question is how many companies will be ready to meet the new regulation when it is introduced?

Many predictions are suggesting that the adoption of AI and automation will place many of us out of work in the future. On the plus side this will in theory free us to live a life of leisure, on the negative side how are you going to pay for food, entertainment, holidays, etc.?

In 1797 Thomas Paine, an 18th-century radical, proposed paying all 21-year-olds a £15 grant funded through a tax on landowners. This was the first time the idea of a Universal Basic Income (UBI) was proposed. More recently with the rise of AI and increased automation the concept of a Universal Basic Income is now being seriously considered with a number of trials taking place in the Netherlands, Italy and Finland. Even in the UK, the Scottish government is planning pilot schemes in Glasgow and Fife. It is also notable that political leaders in France and South Korea have campaigned for the idea and even Elon Musk has advocated the need for UBI.

So what is a universal income? In a nutshell it is an unconditional income paid by the government to all citizens, whether or not they are in work. In Finland 2,000 unemployed people between the ages of 25 and 58 will receive a guaranteed income of €560 a month. It is hoped that this will encourage individuals to start their own businesses but also make it easier for them to take short term jobs which otherwise may affect their benefits. In the UK a figure of around £4,000 a year is being mooted which would cost the government £18bn annually. It should be noted that the aim of providing funding is not to take someone out of poverty, but to give them the flexibility to retrain or the breathing room to wait to take a job that has prospects rather than being forced into taking the first vacancy that comes along. Much larger sums of money would be required to support a workforce that had been replaced by automation.

So is UBI a magic bullet for a struggling welfare state? On the one hand welfare systems are designed as a contributory system of unemployment insurance, in which workers put in during the good times, and take out during temporary periods of unemployment. The reality, however, is that a large portion of welfare spending currently goes on permanently supporting people in jobs that do not pay enough to support their families. This leads to highly complex systems which are expensive to manage. So UBI offers the opportunity to simplify things in the shorter term. The problem is that the labour market is getting more precarious with less full-time employment and more low-paid, low-skilled work. With more automation some predictions are that around half the current workforce will not have employment at all in the future.

So can we afford UBI? It is not cheap to pay every citizen an unconditional income. Polls show that 64% of people in Europe are supportive of a UBI but they were not asked the question would they pay much higher taxes to support this? Notably in Switzerland where the idea of higher taxation was proposed to cover UBI the population voted 78% against the idea. A problem is that if only half the current workforce is still working how could they possibly pay enough tax to pay for all the people who are not working? One suggestion has been the introduction of a robot tax so that companies that use robots to replace workers need to pay a tax for the robot worker. Of course all of this is assuming that new jobs will not be created – notably over the years many jobs have become redundant through the introduction of technology only to be replaced by new jobs. So will we need to work in the future? I think that the short answer is I’m afraid so.

This may be a question you will be asking yourselves in coming years. A number of years ago the OBD port was introduced on cars to give access to the engine fault codes for diagnostics as well as emissions information in order to perform emissions checks. The key thing was that this was provided in an open sense in order to allow third parties to access and use the information. The world is, however, changing and recent cyber attacks have highlighted the vulnerability of having such open interfaces to Cyber Physical Systems. This is particularly a problem as we move towards autonomous cars. As a consequence many car companies such as BMW are announcing that they plan is to progressively close access to the OBD port on the grounds of security and safety. In the first instance this will be to close access to the port as the car is driven but increasingly there are concerns about connectivity and the implications that this may have on warranty.

This presents a problem as a whole service industry has grown up around utilising the data from the OBD port for fleet management, remote diagnostics, insurance and eco-driving. The access to the OBD is actually dictated by EU competition laws and car companies need to allow access to data for garages to allow independent repairers to work on cars. So in future manufacturers still need to allow access to data. So how can this be done?
Looking to the future the aim is to provide an open, standardised, independent and normalised data set via connectivity leading to the concept of the “Extended Vehicle”. OEMs can provide the secured transmission of data to a platform which is separate from the car. The data in this platform can then be made available to 3rd parties, but at a cost. Of course, this opens up opportunities for major players in handling data such as Apple, Google and Amazon. With this comes the question of who owns the data? Certainly, the concept of the extended vehicle raises issues on who has access to the data and also the privacy of the vehicle owner. Additionally, however, the consumer must have the freedom to choose service providers, the parts that are used and also any new applications that may be installed on the vehicle. This will require careful regulation to avoid abuses.

So what is industry doing? Already there are industrial initiatives trying to address these issues. One such initiative is the CARUSO Dataplace. This is being led by a consortium of part manufacturers, distributors and repairers represented by the association of European Automotive Suppliers (CLEPA). This aims to define an open, neutral, interoperable, standardised and secured data cloud and service platform for the automotive sector.

So the future looks very much more connected with vehicle data being held in a large database probably being controlled by the OEMs that by definition needs to be accessible by independent aftermarket providers. This raises an interesting question of what would the OEMs learn from the service providers who access the data and how independent will service providers remain in future?

I have just read that in the UK more than half of jobs will be automated by the early 2030s. This is an interesting story for me as I remember being told at school in the 1970’s that I need to consider leisure activities as it was likely that by the time I was working everybody would only be working 3 days a week. Sadly, this prediction was wrong and so I wonder how true the latest predictions are?

The story is not new. In the UK the Luddites feared machines which led to violent opposition to the industrial revolution in the early 1800s. So will the 4th Industrial Revolution, with the rise of robots, artificial intelligence and related CPS technologies be equivalently painful? It is true that more and more workers, industry leaders and governments are getting concerned and estimates of the impact of automation on jobs by analysts such as PwC - automation will take 40 per cent of US jobs by 2030, and the Bank of England - 15 million jobs in the UK may go (half the current working population!) are of course catching the headlines.

So is automation and AI really a threat or an opportunity? The major hope, as was highlighted to me in the 1970s, is that robots will do the boring tasks to leave us to focus on the more interesting challenges. Indeed looking at all the functionality I use today I am far more efficient and I do not waste as much time writing letters, sending faxes, buying things, booking flights, hotels, etc. So my life has improved immeasurably through technology and I am still extremely busy. The real threat occurs if the technology is monopolised by a few companies who become dominant and take all the profits. There are, however, more positive noises coming from society about automation. ARM found 61% of people questioned thought that AI and more automation will improve society rather than destroy it. There is also some evidence that the introduction of AI is creating new jobs within companies, mainly with the rush in companies to exploit AI for new functionalities.

So should we be embracing AI and robots to improve our lives? Certainly at a national level adopting AI and Robotics is on many national agendas in order to be more competitive so the revolution is coming whether we like it or not. The question is really what can be done to ease the transition for society to increased automation. Bill Gates has advocated that robots should pay taxes in order to replace the lost income from workers who would normally pay taxes. This seems like a good thing, particularly if we move to a single universal income as is being proposed in some countries. On a more negative note Elon Musk has suggested that AI may lead to world war three much like a Terminator film but perhaps we are getting confused here between fiction and reality.

Based on my experience from the 1970s (a time when we were encouraged to build nuclear shelters and were trained for a 4 minute warning) I think we will still have jobs to do but they will be different jobs. It may well be that we work more from home in the future. Of course this will have social implications – the annual Christmas party is likely to suffer - but I think we have enough imagination to come up with new things to do with more free time if we are offered it.
So what does the future hold? I asked Alexa but she did not seem to know ….

First of all a Happy New Year to everyone! I am just back from a winter holiday in Mexico and I am sat in an airport wondering what would I should write about. Of course every aspect of air travel involves CPS whether it is the aircraft themselves, the Air Traffic Management Systems, the baggage handling systems, and airport services used by passengers. However, were you aware that British Airways is using remote controlled tugs to push back aircraft? The electric-powered Mototoks are being used at Heathrow Terminal 5 and have already moved 100,000 passengers.

So what could be the next “big thing” in the industry. While not a new concept (I remember looking at prototype electric two seater aircraft around 10 years ago) there is now a buzz in the industry on whether the future will see much larger electric aircraft for short haul flights. Here there is some promise in developing battery-propelled aircraft that can allow flights under two hours. This would certainly cut emissions and noise which are both key concerns in the aircraft industry. With the increasing debate on pollution, for instance whether diesel cars should be banned in cities, it is no surprise that the industry is looking for new ideas. As we have seen with the car industry the disruptions in technology are coming from new companies such as Tesla with the development of electric cars and trucks. This is driving much higher battery capacities while at the same time lowering the cost of battery technologies. So if it is possible to build an electric lorry that has an 800km range, the question is can the same be done for aircraft?

Airbus via VoltAir are working on the E-Fan aircraft, a two seat trainer, with LiPo batteries in the wings feeding 20kW electric fans. The goal is a 40-minute flight which would give the aircraft a range of 100km. However, this is far from an aircraft that could be used for mass transportation of passengers. It is no surprise, however, that again the disruptive vision of the future is being pursued by non-traditionally companies. EasyJet, for instance is collaborating with Wright Electric to develop and build a 220 seater electric short haul aircraft that could be flying in the next decade. According to the figures this would be 50 per cent quieter and 10 per cent cheaper for airlines to buy and operate than a traditional aircraft. The maximum range of this though is only 335 miles (539km), so London to Paris is possible, but a flight across Europe is still a dream.

So what are the challenges?

Commercial success depends on a number of factors. Foremost there is a need to improve battery technology and improve aerodynamic efficiency to enable much longer flights to make an electric aircraft a practical proposition. Notably as aircraft are very expensive they are operated in the air as much as possible to generate revenue. A short haul aircraft will make several flights a day and it is not uncommon for aircraft to operate for 18 hours a day. There is also a public perception challenge. Passengers are comfortable having two large jet engines under the wings. They may be less comfortable being confronted by an aircraft with many smaller electric motors attached to the wings. Finally, in order to switch from gas turbine engines to electric aircraft there is also a need for infrastructure to charge aircraft and whole new maintenance procedures. So it is likely to be some time before we “fan” off on holiday in the future!

Already we are seeing a proliferation of Advanced Driver Assist Systems (ADAS) being rolled out onto the market. A major step forward though that would have truly profound implications on society would be the adoption of fully autonomous cars (so-called level 4). At this point there is no need for a driver and the car could operate with or without a driver on board. This opens up a number of interesting scenarios that could radically change our society. With a totally autonomous car occupants would have the option of doing other productive tasks while travelling. This would allow passengers to use their car as an office further blurring the distinction between home and work place, it could allow them to surf the web, watch a movie, study and gain new knowledge, interact with friends and relatives or even sleep. How this new free time is used is up to the occupant.
So there are lots of other more productive things that we could do and no doubt there will be no shortage of new business opportunities to sell services tailored to commuter preferences. However, being engrossed in other activities naturally impairs the ability of occupants to react and take control of the car if there is a failure. There is thus a need to consider whether changes in the way occupants use their time within a car will result in a shift to fully autonomous control even in the event of an accident.
The reduced risk of accidents through the use of fully autonomous cars could also result in changes in the law. For instance if it could be proved to be safe speed limits could be increased and traffic could move faster. Likewise if accidents are a very rare event in the future car manufacturers may well start to design out current safety features designed to minimise injury in an impact with a consequent reduction in the weight of the car and an improvement in efficiency.
A key question is whether fully autonomous cars will lead to a future of mobility solutions? If a car does not need a driver then there is also the option of the car operating by itself without any occupants. This facility could be further used to save time and be more efficient. A car could drop off its passengers and then go and park itself. It could also come and pick up passengers form work and home. At this point the need to have car parked at the office or at home is no longer necessary as long as it is scheduled to arrive at suitable times. This could significantly reduce congestion or allow the car to be shared with other users when not needed by the owner. The car could also go and pick up children from school so that parents do not have to leave work. The car could also take itself for maintenance and service. Here there could be potential benefits in fuel economy and emissions as many cars are operated with dirty air filters and incorrect tyre pressures that result in loss of efficiency.
In terms of logistics the future may bring autonomous deliveries to the door. The most likely scenario initially is that people would need to leave their houses and transfer goods to their house from the autonomous vehicle. However, people may well send their cars to pick up goods and elect to meet their shopping at appropriate places such as a car park or a place of work.
The removal of the need for a driver could have significant and profound negative impacts on society. There would be no need for taxi drivers (indeed many taxis are shared by drivers – as one driver rests another driver utilises the taxi) so removal of the need for a driver may allow an operator to actually replace two drivers with consequent savings of two salaries, that would quickly cover the cost of the autonomous taxi, and result in greater loss of employment. Likewise bus drivers would also face unemployment. Looking to the road freight industry the introduction of autonomous trucks and vans could result in massive unemployment across Europe.
So the implications of fully autonomous cars are wide ranging impacting on our way of life, how we use our time and our employment prospects for the future. A revolution is coming but are we ready for it?

In a future world with autonomous traffic it needs to be remembered, that not all vehicles will be autonomous. There will be a mix of autonomous and human driven vehicles and also pedestrians. Humans are naturally unpredictable and this lack of predictability presents huge challenges to autonomous systems and is likely to be the main cause of accidents in the future. From a social standpoint drivers and pedestrians may well need to adapt the way they drive and interact with roads and traffic in order not to disrupt the autonomous flow of vehicles. A question is how well will drivers integrate with a traffic flow of autonomous cars? This could be daunting for drivers who are not used to interacting with autonomous vehicles and could well be frustrating to some drivers who wish to have the freedom to get from A to B using their own driving style or via their own preferred route. It should be noted that autonomous cars will always follow the law whereas humans may elect to break the law by speeding, jumping a light or even ignoring road directions. It is a fact of life that autonomous cars must be able to react appropriately to bad human driving.
Humans may also not do what an autonomous car may anticipate. A classic problem is a driver deciding to stop unexpectedly in the middle of the road either intentionally or due to a sudden break down. Systems must be designed with high safety requirements and human-friendly behaviour which accounts for such unpredictable actions. These needs should be addressed both at a low level (e.g. guaranteed reliable sensors for people, car and obstacle detection), at a systems level (e.g. new algorithms for people tracking and new human-robot interfaces) and finally through systems for validation and certification.
So what will be the impact on traffic flow and society as a whole as we get more autonomous cars on the road? A key consequence will be control of traffic flow. This will naturally reduce congestion and should produce more predictable journey times. Drivers in future should be happier because they are not stuck in traffic, they should be less frustrated and stressed and should also benefit financially because they are driving more efficiently. More consistent traffic flow will lead to more predictable journey times with better on-time arrival for work and meetings.
However, what concerns me is what happens if I get up late and need to get to work quickly? I’m afraid if traffic flow becomes regulated there will be no option to get up late in future as it will not be possible to drive faster. So in future it looks like we will need to always ensure that we set off on time – or work from home of course…

The introduction of autonomous vehicles will occur in an incremental fashion as the regulatory environment develops. It will take time for the existing regulatory system to adapt to the introduction of autonomous transport. This is complicated by the uncertainty on what should be regulated and the risk that regulation may lock-in one pathway to automation over a potentially better one. While needed, early regulatory action carries risks as well. Some regulatory flexibility seems desirable, for instance putting in place regulation to cover certain use cases, e.g. to allow low speed urban operation or motorway platooning, in advance of defining a blanket set of rules for autonomous driving in the future. Policy makers should account for this uncertainty and be ready to adapt to changes.
There are a number of legislative hurdles that need to be addressed before a 100% autonomous car in Europe is possible. One problem is the non-uniformity in laws to allow data to be obtained from infrastructure; in some European countries law does not allow communication between traffic lights and cars. Another issue deals with information on the speed of cars which could be potentially also be used for identifying speeding. The shifting the driving task from humans to machines will require changes in the insurance industry. Issues that need to be dealt with are how in the case of accidents responsibility should be apportioned among suppliers and sub-suppliers and what should victims have to do to get support for their loss and/or recovery (without long battles in court). Until the liability problem is solved, the manufacturers and designers of autonomous vehicles are hesitant to bring their products to the market. There is thus a need for a clear framework for legally handling the consequences of accidents with autonomous cars. However, adjudication methods have yet to be developed.
So we need flexibility in regulation, harmonisation across Europe and clarification from the insurance industry. A lot to ask but all are key requirements if we are to see the successful roll out of autonomous cars on our roads.

From a technological standpoint the main challenges for CPS applications for autonomous driving vehicles concern the development of autonomous navigation, map building and localisation and the ability to operate in a dynamic environment with a mixture of autonomous and non-autonomous vehicles and pedestrians. This requires reliable sensing technology in combination with viable deployment strategies.
However, technology is one thing, there is also a need for societal acceptance and here the main requirement is to develop trust in autonomous cars. This can only be delivered through long-term successful demonstration of safe autonomous functionality to establish confidence. A key problem is that inevitably accidents will still happen. Here there is a need to educate people as it is clear that the media will pay particular attention to crashes of autonomous cars while not considering the overall improvements in safety provided by increasingly autonomous traffic. There are also other more practical barriers. For instance, there is a need for insurance companies to agree on an approach to apportioning liability before full autonomous functionality can be rolled out onto Europe’s roads.
Security will also become an increasingly important issue as autonomous cars become highly interconnected. Systems need to be secure (no outside entity should be able to break into the system and cause an accident) but also there is a need for the vehicle to fail-safe even in the presence of a security breach. The obvious action in this case is to disable the engine and pull the vehicle over to the side of the road to a place of safety. Here designers will need to wrestle with the problems of providing security, privacy, and fail safe design to operate in presence of security breaches.
With recent ransomware events the general population is becoming increasingly aware of the dangers of cyber-attack. Of course from a social standpoint there are likely to be concerns about entities disrupting traffic externally causing deliberate gridlock. In the future citizens may well believe that gridlock is a result of autonomous system failure or failure in intelligent traffic management systems, or even terrorist action. How will we know? So in the future it will be important to reassure drivers on their journeys that the traffic network is operating correctly and safely at all times – but how can we get them to trust us?

The primary interest of the automotive industry in introducing autonomous driving is to improve road safety. An immediate benefit from this is reduction in deaths. However, there will also be a reduction in serious injuries reducing suffering of citizens at an individual level while at the same time saving medical and insurance costs which will benefit society as a whole.
Indeed safety of autonomous cars is the main driver for their market introduction. As most crashes involve human error, and autonomous operation can reduce or eliminate these errors, exploitation of CPS is expected to benefit road safety. Although there have been notable high-profile failures with autonomous cars crashing from companies working at the “bleeding edge” of the technology the results from early cars are promising. The key problem appears to be that it is possible to design for known scenarios but unfortunately it is not possible to design for all the unforeseen scenarios that occur in the real world.
Also looking to the future as autonomous technologies become more common, new types of crashes may emerge, for instance crashes resulting from the car handing control back to the driver unexpectedly or from mixing autonomous and conventional vehicles on the roads. If an accident is imminent there may well be cases where handing back control to the human is the wrong course of action. Here there are very difficult decisions to be made on behalf of humans about whether they are capable of reacting in the most appropriate way. This loss of authority may well be totally unacceptable to citizens even it means that they will suffer more serious injuries or death as a result of a crash if control is returned to them at the last moment.
So we will get safer roads, but there is a lot of work to be done on understanding how to deal with unusual situations on the fly and consider when to allow drivers authority to take control.

Cyber-Physical Systems (CPS) are expected to revolutionise the transport sector with an impact on safety, emissions and the efficiency of transport of persons and goods across Europe. Europe has several world leading automotive companies and some of the largest logistics companies with extensive supply chains. Car manufacturers are already announcing the first applications of autonomous vehicles and the levels of automation will increase as the regulatory environment develops.

A key argument for autonomous cars is that fatalities and injuries on Europe’s roads will be reduced and European citizens will enjoy continued mobility as they age. Most car manufacturers and technology firms are introducing Advanced Driver Assistance Systems (ADAS) with features such as supporting the driver to hold a safe speed in relation to the vehicle in front (adaptive cruise control), to maintain position in the lane (lane departure warning) and to park (park assist).

So what does the future hold? It is expected that commercial production of highly automated passenger vehicles will start this year with a wide range of models being on the market by 2030, some of which are likely to be self-driving. However, the introduction of autonomous vehicles will be incremental; step-by-step with increasing levels of automation. This is because the vehicle manufacturers will steadily increase the technological level of automation in their products in concordance with the adaption of the regulatory environment. But will cars be totally autonomous in the future? Interestingly, all the future market predictions at present do not see totally self-driving cars being on the road by 2030 so it seems there are still a number of barriers we need to address first before we can sit back and let our cars drive themselves.