1、 ISSUES AND OPTIONS FOR LOCAL AUTHORITIES Wellington House, 40-50 Wellington Street, Leeds LS1 2DE 0113 251 7204 infourbantransportgroup.org ISSUES AND OPTIONS FOR TRANSPORT AUTHORITIES ON CONNECTED AND AUTONOMOUS VEHICLES AUTOMATIC FOR THE PEOPLE? The Urban Transport Group represents the seven stra

2、tegic transport bodies which between them serve more than twenty million people in Greater Manchester (Transport for Greater Manchester), Liverpool City Region (Merseytravel), London (Transport for London), Sheffield City Region (South Yorkshire Passenger Transport Executive), Tyne and Wear (Nexus),

3、 West Midlands (Transport for West Midlands) and West Yorkshire (West Yorkshire Combined Authority). The Urban Transport Group is also a wider professional network with associate members in Strathclyde, Bristol and the West of England, Tees Valley, Nottingham and Northern Ireland. Automatic for the

4、people? Report authors: Rebecca Fuller, with thanks to Clare Linton and Jonathan Bray.March 2020 2 CONTENTS Contents 1 Introduction .4 2 Introducing Connected and Autonomous Vehicles .6 3 UK priorities and policy position on Connected and Autonomous Vehicles .12 4 Exploring the case for Connected an

5、d Autonomous Vehicles .18 5 Key issues for transport authorities on Connected and Autonomous Vehicles .26 6 Options for transport authorities on Connected and Autonomous Vehicles .35 7 Conclusions and recommendations .40 8 References . 44 3 INTRODUCTION Connected and autonomous vehicles (CAVs) can b

6、e considered a Marmite issue. People tend to either instinctively love or hate the idea of fully autonomous vehicles with little middle ground. Reports on the subject are often equally polarised, taking either an evangelical or a doom-laden tone as they imagine the utopia or dystopia they will event

7、ually unleash. Some imagination is required in either case given that nobody can say for certain exactly how CAVs will evolve, at what pace and with what consequences. Envisaging a future with CAVs involves making a number of assumptions about how people will respond to them; at what level of autono

8、my and connectivity they will predominantly operate and in what circumstances; what the applications might be for different modes of transport; and how soon (if ever) they will come to dominate the transport mix. These assumptions have implications for transport policy, networks and infrastructure b

9、ut also for a wide range of other policy areas, from public health to urban planning. This helps to explain why CAVs have also been described as a wicked problem1. A wicked problem is one that is difficult (or impossible) to solve because of incomplete, contradictory and changing requirements. It in

10、volves many stakeholders with differing views, a large potential economic impact, no determinable stopping point and complex interdependencies. Wicked problems are not bad but are extremely challenging for policy makers. The CAV industry also increasingly recognises the challenges that mass deployme

11、nt of the technology presents, requiring government approval, public trust, brand marketing, the ability to manufacture at scale and the technical knowhow to manage a fleet. 2 Whilst this report cannot give definitive answers as to what the future trajectory for the development of CAVs will be, it d

12、oes aim to provide an objective framework for city regions to think about CAVs, their implications for wider priorities and the approaches they might therefore take. In doing so, this report seeks to do three things that set it apart from the many other reports there are on CAVs. Firstly, it specifi

13、cally looks at CAVs from the perspective of city region transport authorities in the context of their wider objectives and responsibilities, rather than considering CAVs in isolation from wider public policy considerations. Secondly, it does not solely focus on the implications of an as yet hypothet

14、ical end state where the vast majority of vehicles are fully autonomous. Instead, it recognises that vehicles are increasingly connected and have begun to operate with features which have a degree of autonomy. We have therefore already embarked on a CAVs trajectory. 4Automatic for the people? 5Intro

15、duction 1 Regardless of whether this trajectory ends in a future where all vehicles are fully autonomous, there are implications in the here and now for transport authorities. This report takes the view that CAVs are, and will continue to be, a moving target with live implications at every twist and

16、 turn rather than a single leap to a pre-determined end state on which all thinking and policy making should focus on accommodating. Thirdly, whilst many other reports on CAVs focus exclusively on how cars might become fully autonomous, this report looks at a wider range of vehicles including buses

17、and public service vehicles. It also looks at the implications of CAV technologies for the roads these vehicles will be travelling on. The report begins by providing clarification on what we mean by CAVs before assessing what stage the technology is at and where it sits in terms of UK government pol

18、icy and priorities. The report then analyses the four key areas often highlighted as representing the main potential contribution of CAVs (safety, economic, social and environmental benefits) before examining six challenges of particular relevance to transport authorities. Consideration is then give

19、n to the options open to transport authorities in terms of how they might approach the emergence of CAVs. It features a number of international examples highlighting a range of approaches from laissez faire to acting as a guiding hand. Finally the report suggests what transport authorities can do no

20、w to respond effectively to the development of CAVs and what actions national government should take to enable them to do so. 6Automatic for the people? Level 0 No Automation Level 1 Driver Assistance Level 2 Partial Automation Level 3 Conditional Automation Level 4 High Automation Level 5 Full Auto

21、mation This is identical to Level 4 except that the driving automation features are not limited by an operational design domain. Instead they are capable of performing all driving functions in all situations that a human driver could. Hands on Hands on Hands on Eyes on Eyes on Eyes on Eyes offHands

22、off Hands offEyes off Eyes off (temporary)Hands off (temporary) INTRODUCING CONNECTED AND AUTONOMOUS VEHICLES Defining Connected and Autonomous Vehicles are automated in some way that is, they can perform some or all functions without driver input. Vehicles can have varying degrees of connectivity a

23、nd automation. In respect of automation, the most widely used model is that developed by SAE which describes six levels of vehicle automation3: For the purposes of this report, we define connected and autonomous vehicles as: those which use an external network connection to communicate in some way e

24、ither to the driver, other vehicles, roadside infrastructure or the cloud, or to any combination (or all) of these and 7Introducing Connected and Autonomous Vehicles The human driver performs all aspects of all driving tasks, even when these are enhanced by warning or intervention systems. The drive

25、r assistance features can carry out either the steering or acceleration/deceleration. The driver assistance features can carry out both steering and acceleration/deceleration. The driving automation features can perform all driving tasks but a human fallback-ready user is expected to respond appropr

26、iately to a request to intervene. The fallback-ready user must be receptive to a handover request or to an evident system failure, but is not expected to monitor the driving environment. The driving automation features can perform all the driving tasks within their operational design domain (for exa

27、mple, motorways only). There is no expectation that the human user will respond to a request to intervene. If the limits of the system are exceeded, the system will put the vehicle into a minimal risk condition, such as a safe stop. This is identical to Level 4 except that the driving automation fea

28、tures are not limited by an operational design domain. Instead they are capable of performing all driving functions in all situations that a human driver could. 2 8Automatic for the people? Connected and autonomous technology has potential applications for a wide range of vehicle types from cars to

29、buses and from road sweepers to HGVs. What stage is CAV technology at? Connectivity features have been built into road vehicles for some time. This could be a connection to a phone, Bluetooth, GPS, sat nav or the internet, via an internal SIM, allowing the vehicle to remain online at all times4. It

30、is estimated that there are over three million vehicles in the UK with internet connectivity5. Most are single vehicles connected to an information cloud. Current connectivity features include smartphone integration, roadside assistance, parking apps, remote diagnostics and voice commands6. The next

31、 level of connected vehicles are still in their infancy but would eventually be able to communicate with one another, with road infrastructure and even across national borders7 with the goal of improving safety and efficiency. Developments in AI and machine learning could see connected vehicles lear

32、ning from their collective experiences. Some autonomous features are common in the vehicles of today, including automatic starters, gearboxes and wipers; emergency braking; cruise control; lane assist; and park assist. It is interesting to note that, outside of the road environment, automation/drive

33、rless operation is relatively common on metro networks around the world. UITP reports8 that globally 25 cities run automated metro lines, with the first dating from 1981. The trains require no driver in the front cabin or staff assigned to a specific train. Vehicles with more autonomous features are

34、, however, beginning to appear on our streets as part of real-world trials of the technology. The recently completed DRIVEN project, for example, tested automated vehicles on the streets of Oxford on a daily basis9 as well as demonstrating autonomous fleet driving in Londons complex urban environmen

35、t10. Meanwhile, robot food11 and package12 delivery pods are successfully navigating the streets of Milton Keynes (albeit monitored by human operators who can step in if required). Whilst there are many more trials and demonstrations planned and underway, we are still a long way away from regularly

36、seeing Level 3-Level 5 autonomous vehicles on our streets. Developers of self-driving cars, for example, have suffered setbacks as they attempt to refine the technology, some with tragic consequences for car occupants and other road users alike. Furthermore, trials and demonstrations are an entirely

37、 different prospect to large-scale deployment. James Farley, Fords President of New Businesses, Technology and Strategy has commented that the development of self-driving systems is necessary, but not enough to build a business on. He talks about the continuing uncertainty around the manufacturabili

38、ty of the sensing systems, the regulatory environment and the scaling of the customer facing business.13 Mindful of the complexities posed by urban environments, CAV developers are increasingly embracing operational design domains14 and geo-fencing. This involves focusing on more predictable, constr

39、ained tasks and environments that the vehicles can handle now, or soon (for example, running on segregated portions of the highway, on set routes between two points or in defined spaces like airports or university campuses). 9Introducing Connected and Autonomous Vehicles CAV testing on campus Salfor

40、d Transport for Greater Manchester is working in partnership with the University of Salford to test a 15-seater Level 4 autonomous vehicle in a real-world environment around the private roads of the University campus. Current vehicle registration requirements (such as the need for the driver to have

41、 a clear view and for a steering wheel) mean it cannot yet be used on public roads. Eventually, the project plans to link the vehicles in to the wider public transport network, connecting people to rail, bus and tram hubs as well as to the Universitys second campus at Media City. Even in the constra

42、ined environment of the University campus, a number of practical challenges have emerged. The team has discovered that the vehicles Lidar system (which acts as eyes) cannot function in horizontal rain or in snow, both of which cause the vehicle to stop. Low hanging branches also stop the vehicle as

43、it is unable to determine what the obstruction is made of and how likely it is to cause damage. A bird landing on a rooftop antennae also impeded the vehicles progress. These small examples highlight the huge challenges CAVs will face when operating in a complex streetscape. The Level 4 autonomous v

44、ehicle on campus. Credit: Rebecca Fuller The Government Office for Science predicts that Automation could plausibly challenge the transport system status quo from the 2030s onwards, for trains and buses but also, potentially for private car-based transport15. Buses, trams and trains are thought of a

45、s natural candidates for early automation given they typically follow predictable routes and can use infrastructure (like bus lanes) designed to minimise encounters with conventional traffic. 10Automatic for the people? Europes first full-sized autonomous bus fleet Fife/Edinburgh Europes first full-

46、sized autonomous passenger bus fleet trial is set to launch this year, serving up to 10,000 passengers per week between Fife and Edinburgh across the Forth Road Bridge, plus on-road and hard shoulder running16. Since 2018 the bridge has been designated as a Public Transport Corridor, with access to

47、motor vehicles (other than buses and taxis) restricted17. Ferrytoll Park and Ride in Fife and Edinburgh Park train and tram interchange sit at either end of the route18, offering connections through to the wider transport network. Each bus will be able to carry up to 42 people and, whilst it will operate at Level 4 automation, a driver will remain on board in line with regulations. Autonomous systems developed for the trial are anticipated to provide spin-offs th