Self-Driving Event Report: BSI Webinar, November 2024
The BSI webinar, “Self-driving vehicles: what’s ahead of us?”, on 13 November, set out to answer as many of your frequently asked questions (FAQs) as possible. How did we do? Judge for yourselves…
Following a brief intro by Nick Fleming, Transport Standards Director at the UK national standards body, Marty Zekas, of the government’s Centre for Connected and Autonomous Vehicles (CCAV), got the ball rolling by giving a detailed update on the Automated Vehicles (AV) Act.
Formerly Bill Manager for the Act, and now leading the implementation program, he reflected on the huge progress made since the Law Commissions began their review of the legal landscape for AVs back in 2018, and then outlined some key areas for the secondary legislation to come.
“The Act covers four main areas – safety, liability, misleading marketing, and the licensing of automated passenger services,” he explained. “It addresses which entity is responsible for oversight of the self-driving vehicle when it’s in self-driving mode.
“If there’s a user-in-charge (UIC), who’s ultimately responsible, and also operator licensing for situations with no human on board, ensuring that we’ve got responsible operators overseeing safe operations. Then we come to incident investigation measures and in-use regulation, ensuring that vehicles continue to meet the safety standards.
“The UK is also working at UN level to harmonise international standards for self-driving vehicles. These are expected to be finalised by June 2026, and come into force in January 2027, which will align quite closely with our regulatory timeline.”
CCAV have 17 workstreams covering all aspects of AV deployment, from technical performance regarding the dynamic driving task, user interactions, service monitoring and virtual testing, to safety case approaches, the requirements on manufacturers, cyber security, and changes to the highway code.
“A major part of this is the work we do with BSI to set important standards to ensure a smooth customer journey,” said Zekas.
Then came the main event – a panel session moderated by yours truly featuring Oliver Howes, International Regulatory Lead at Oxa, Siddartha Khastgir, Director of Partners for Automated Vehicle Education United Kingdom (PAVE UK), Brian Wong, Partner at law firm Burges Salmon, and David Wong, Head of Technology and Innovation at The Society of Motor Manufacturers & Traders (SMMT).
We covered a lot of ground – from the remaining economic, legal and technical challenges, through myth-busting and public trust, to – whisper it – a sizeable shift away from private car ownership.
We started, however, with the ‘what, where, when and why’ of early use cases, so, as a UK company pioneering deployment around the world, we went to Oliver Howes first.
Early use cases
“At Oxa, we focus on a few different high-level use cases – industrial logistics, passenger transit and asset monitoring,” he said. “Then, there’s a whole load of sub-use cases – from HGVs moving goods from hub to hub, and small guided vehicles operating in warehouses, to passenger shuttling for demand-responsive transport, and buggies operating off-highway, doing jobs which would put humans in danger.
“We are developing our software as a universal solution that can be deployed on any vehicle platform. When it comes to passenger transit, we really want to focus on the larger issues. Shuttling allows you to replace many cars with one, improving emissions and congestion. We’re currently supporting live deployments in the US, in California and in Florida, with our partner Beep.”
David Wong divided early use cases into three ‘buckets’. First, automated driving features in UIC vehicles – for example, a level three automated lane keeping system (ALKS) or automated valet parking. Second, automated passenger services – ride-hailing taxis, shuttles, buses or coaches. Third, vehicles for industrial operations – logistics and delivery – first mile, last mile, middle mile or off-road, for example, baggage handling dollies at airports or automated mining vehicles.
“Some of these are already commercially deployed, just perhaps not in large numbers,” he said. “Passenger cars fitted with ALKS, the world’s first automated driving feature (under United Nations technical regulation 157), have seen early deployment in very small numbers in Germany, California and Nevada.
“There are already ride hailing vehicles in San Francisco, Phoenix and LA, and shuttles running in Florida and California, and a commercial service in Wuhan, China, which I understand has been very successful. For the third application, we can already see pilot deployment of automated middle mile and long-haul trucking services in America, particularly in Texas.”
Brian Wong noted that, by the time the first authorised, approved, certified and licensed AV takes to a UK public road, it will probably be the most tested and validated vehicle in the country.
Returning to the ‘why self-driving?’ question, Khastgir said: “The UK has good road safety, but not great. For the last 10 years, we’ve had 1 ,700 deaths on our roads and it has stagnated at that level.
“The Secretary of State, Louise Haigh, told the Transport Select Committee that road safety should be considered a pandemic,” he said. “So, it’s important to appreciate that technology, ADAS and automation, will provide near term benefits.”
Remaining challenges
Next question: The remaining challenges – economic, legal and technical (everything apart from public trust, which we’ll come to in a moment).
Talking about last year’s Mi-Link project, the UK’s 1st fully electric autonomous bus trial, John Birtwistle of First Bus asserted that the most significant hurdle is now financial.
Brian Wong said: “I know John very well, and we should listen to him, because his company has moved hundreds of millions of people. Those of us who have been around this industry for a while know that a lot of roadmaps have been produced to bring connected and automated mobility (CAM) to life. On the regulatory side, the UK has always performed extremely well – all the work by the Law Commissions, CCAV and the standards bodies.
“In 2014, when Burges Salmon started getting involved in driverless cars, via a project with AXA, working groups were sometimes baffled as to why lawyers and insurers were there. We made the case. We explained what needed to happen. One by one, the systemic blockers are being removed and the requested frameworks are being put in place.
“The Automated Vehicles Act will be underpinned by a lot of secondary legislation, providing much more transparency than in some other countries. Even before that comes in, there’s a lot the likes of Oxa and Wayve can do using the UK’s existing trialling guidance.
“We forget sometimes that the Docklands Light Railway (DLR) has run in London since 1987, carrying 90 million passengers a year on a largely automated basis. The Heathrow pod has been running for well over a decade, carrying thousands of people a day. These are the ancestors of automated vehicles.”
Praising CCAV and the DfT for doing ‘a fantastic job’ on the primary legislation, Khastgir warned: “The proof of the pudding is actually in the secondary legislation, what Marty was talking about for next year. That’s where the big challenges lie.
“What is the safety threshold? We don’t yet know what benchmarks governments are expecting. This leads to technical challenges – how do you demonstrate it? Also, we should not kid ourselves – this technology is going to be expensive. Those with deep pockets will achieve SAE level three, level four systems, others will not be able to.”
David Wong pointed to the SMMT’s 2016 framework for understanding the barriers to commercialising and deploying self-driving vehicles. “Eight years on, the work remains very much intact,” he said. “There are four challenges. First is the engineering challenge. Put enough time, money and brain power in and somebody will crack that.
“Second is the regulatory challenge. The sooner we have the secondary legislation the better. Third, is the public acceptance issue – bringing people on the journey. And fourth, probably the most intractable in relative terms, is the economic case. It’s not just about the hardware or the total cost of a vehicle, it’s about finding the right business model to deliver value.”
On the economics, Howes emphasised that the high capital expenditure required to buy these vehicles is offset by reduced operating costs. On the technology side, he said: “There’s a lack of native driverless platforms being designed and developed across the world. Often, we’re still retrofitting existing type approved vehicles. The Ford e-Transit is probably the most accessible platform.”
Trust in self-driving
To set the scene for the consumer confidence section, we highlighted a study raised in advance by a webinar subscriber: Liza Dixon’s 2020 paper comparing the scientific literature on driving automation, claims by OEMs, and media reporting.
Coining the term “Autonowashing” to describe the practice of making unverified or misleading claims which misrepresent the appropriate level of human supervision required, the associated case study related to Tesla.
As head of PAVE UK, with its mission to educate the public, we naturally turned to Professor Khastgir first. “Lisa’s paper is seminal in this space,” he said. “Essentially, it’s a very simple concept – trying to sell something as it is not. If you go back to the cognitive psychology literature, there is enough evidence to say that, in order to build trust in a system, you need to accurately tell the person what it is capable of doing and what it is not capable of doing.
“The concept of absolute safety is a myth. There is no technology in the world, be it your laptop, your phone, the car that you drive today, which is 100% safe. But you can still use the technology in a very safe manner if you use it within the boundaries of its operational design domain (ODD).
“It is now the responsibility of every player in the ecosystem to make sure people understand the true capabilities and limitations of the technology, particularly the difference between ADAS and self-driving. The UK should be really proud of its approach to misleading marketing in the legislation. There’s a role for standards too, both in terms of nomenclature and performance. I like the term ‘thoughtful standardisation’ – standards which add clarity.”
David Wong immediately picked up on this, saying: “For our part, as the industry body, we recognized that importance early on, which is why we developed two things: First, voluntary guidance for our members and car manufacturers regarding advertising self-driving features; and second, working with stakeholders and CCAV to produce a toolkit in preparation for the first passenger cars fitted with self-driving features.
“If you ask a person in the pub, what’s an ODD? What’s a transition demand? Nobody knows. So, we must use plain everyday language. That’s absolutely pivotal, because if the public are confused, and they misuse or abuse a particular technology, and something goes wrong, you risk setting the industry back many years.”
Howes asserted that getting more members of the public hands-on with the tech is the best way to build trust. “There are lots of trials going on with Waymo in the US at the moment, which are really improving the public perception of automated vehicles, but it’s very limited in the UK and Europe,” he said.
Brain Wong reiterated the vital importance of ‘not messing around’ with creative advertising when lives are at stake. As to spurious reporting in the media, our Hyperbolic Headlines strand illustrates the scale of the task.
Ownership implications
With Uber CEO, Dara Khosrowshahi, recently saying that Wayve’s advanced Embodied AI “holds a ton of promise as we work towards a world where modern vehicles are shared, electric and autonomous”, David Wong turned his attention to what self-driving might mean for private vehicle ownership.
“There is a place for shared mobility, particularly in densely populated urban centres,” he said. “But the personal passenger car that is manually driven, or perhaps has a manually driven mode, will still be here for many years to come.
“There is a coming together of megatrends – we used to say ACES (automated, connected, electric and shared) and now we say CAM. But we need to tread this path very carefully, to find a way to be resilient if there’s another pandemic and people again shy away from shared mobility.”
Quickfire Q&A
Now, with time flying, we quickly ran through some live questions from the audience. How will artificial intelligence (AI) improve self-driving? We won’t have self-driving without it, simple as that.
How will self-driving cars interact with cyclists and pedestrians? Just complying with traffic rules doesn’t necessarily mean the behaviour is safe. Self-driving vehicles must behave appropriately for the environment and situation. How long will it be before all vehicles are self-driving? Probably not in your lifetime.
And finally…
Just before we wrapped-up, BSI’s CAM programme lead, Matteo Novati, gave an update on the latest standards work.
“We’ve heard about building trust through awareness and transparency, and about delivering value and efficiency to overcome some of the economic challenges – these are some of the objectives of industry standards,” he said.
“Our CAM program has been running for over five years. We’ve published 10 standards that are openly available, developed thanks to the inputs of more than eighty organisations. Our CAM standards roadmap is one of the key tools that can help the industry prepare for the deployment we’ve been talking about. It signposts the most up-to-date guidance that is available internationally.
“We also heard during the panel discussion about building confidence and trust in the safety of AVs. Our Flex 1888 proposes a harmonized framework for selecting minimal risk maneuvers, when the vehicle has to respond to issues that can prevent the continuation of the driving task. It is based on the assessment of the relative risks, and tries to shed light on how to select the most appropriate achievable risk conditions.
“We’re working with Siddhartha and WMG on a new ODD taxonomy to support the safety assurance process that the UN working group is proposing. We will continue to engage on priority areas, including the harmonisation of measurements and metrics, and the potential standardisation of vehicle behaviours in relation to emergency response services.”
For further info, please see the CAM section of the BSI website.
