If you listen to the government, self-driving vehicles are just a minute away from hitting our roads. It believes the sector could create up to 38,000 jobs and unlock an industry worth £42bn by 2035. In December 2025, the Department for Transport launched a call for evidence on how self-driving vehicles should be safely introduced, announced a passenger piloting scheme for the spring of 2026 and indicated that a full regulatory framework would arrive from the second half of 2027.
So it’s full speed ahead for self-driving, the government believes. And we’ve already got an idea of where those vehicles will be. Milton Keynes has expanded low-speed autonomous shuttle trials, including a route between the station and the city’s Hotel La Tour, while in London Uber and British start-up Wayve say they plan to develop and launch Level 4 self-driving trials timed to coincide with the government’s push for earlier commercial pilots. Level 4 autonomy means the vehicle can handle all driving tasks on its own within specific, pre-defined conditions or areas, without needing a human to take over – but it isn’t designed to work everywhere or in all weather and traffic conditions.
However, for mechanical engineers, the interesting question is not just what autonomy can do. It’s how a new safety regime, new data expectations and new business models will reshape what vehicles are, and what it means to engineer them.
Going slow then speeding up
In many ways, the self-driving revolution is already here. “We have autonomous vehicles on the road today,” says Yousif Al-Ani, principal advanced driver assistance systems engineer at Thatcham Research. “They’re approved under a trial.” But Al-Ani points out that you can’t walk into a car dealership and drive off with a new self-driving vehicle. In other words, the tech exists, but is being used in specific tests and trials with clearly defined operating conditions and many layers of oversight.
Some of that oversight comes from the 2024 Automated Vehicles Act, passed by the House of Commons. It established a self-driving safety framework and defined an “authorised automated vehicle” as one that must be able to travel autonomously “safely and legally”. The act has helped bring autonomous vehicles (AVs) to the point they are at now, but more work will be needed on the legislative side for the vision of the future that people like Elon Musk want to see on our roads. Al-Ani points out that secondary legislation will help establish the detailed rules that decide what evidence is required, what monitoring is mandatory and what happens after an incident if something goes wrong.
Engineers will play their role in helping to develop tests informed by secondary legislation, but it is for that reason that – at least for a while – self-driving vehicles are likely to remain the type we see today. “You will not see a future where you have to stop manufacturing or rethink how vehicles look because it’s going to be a heterogeneous mix for a very long time,” says Siddartha Khastgir, head of safe autonomy at Warwick Manufacturing Group, University of Warwick.
While vehicle designers might want to imagine a world where they can engineer total autonomy, they’ll have to engineer how manual cars, assisted-driving cars, shuttles, delivery bots, cyclists and pedestrians will all share the same streets for a little while longer.
Inside story
One question engineers may have when thinking about the vehicles of the future in a self-driving world is whether the interior of vehicles could change. If people in a vehicle aren’t going to be driving, do they need the things that help them drive cars, such as pedals and a steering wheel? For the moment, the rules require it. “The vehicle needs to have a steering wheel because that’s part of the regulation,” Khastgir says. “It’s not rocket science. It’s just a process we have to follow.”
But even if the technology got better, Khastgir says the steering wheel is likely to persist, partly because it makes authorisation easier. “You have them because it’s easier to get the vehicle approved, because the theory behind this is that the driver could take over if the vehicle can’t work,” he explains. Removing the wheel isn’t just a change in the design. It’s a reshaping of responsibilities: who is in charge when something goes wrong, how occupants interact with the vehicle and how to certify a vehicle with no manual fallback.
Still, the legislation is moving as engineers roll out new self-driving features into vehicles. In 2021, the Department for Transport set out how vehicles fitted with automated lane keeping system technology could be legally defined as self-driving, limited to motorways in slow traffic and speeds up to 37mph. It’s possible to envisage how the same thing could happen for more innovative engineering solutions in self-driving capabilities.
The UK is moving slower than other countries because of different approaches to legislative constraints, according to those in the industry. While nearly half a million trips every week are made in self-driving vehicles operated by Waymo alone in the US, and the equivalent figure is at least a quarter of a million a week in China, elsewhere adoption has been more hesitant.
Some think this is because of the difference in road layouts and natural environments. Ask most people what will stop self-driving cars in Britain and they point to its roads: rain, narrow lanes, parked vans, roundabouts. But Al-Ani thinks this excuse is misplaced. “I don’t think that’s the big challenge, and that’s what everybody hinges on,” he says. “The environment isn’t really the problem. It’s more operational safety and processes that will have to come in in the UK to make everything just work without manual intervention.”
The engineering challenge of self-driving has been largely solved, says Al-Ani, but the issue now is around the need to ensure that rules and regulations – both at a governmental and private-sector level, including insurance – are in place for the transport of the future. Rather than being a hardware problem to solve, this is more of a software challenge. “In the event of a collision, the insurer has to pay out, and then they have the right to subrogate if the automated system was on,” Al-Ani says. “But the problem is the subrogation part; unless an insurer actually has access to data from the vehicle, they’re not going to know whether that system was on or off, and whether they can subrogate. It’s going to be a case-by-case investigation on every collision.”
As Al-Ani puts it, “it’s not the technical problem of ‘Can the car actually drive down the road?’ any more. I think we’re past that”. The bottleneck now is the more mundane aspects such as authorisation, monitoring and incident investigation, which keep a fleet safe as software changes and ensure it is both insurable and roadworthy.
Building trust
Engineers play a role in that because they can reassure the regulators and the public that the technology is ready for real-world rollout. “The important thing for any new technology is for people who experience the technology to be able to build trust,” says Khastgir. “That’s the fundamental requirement to build societal trust in technology.” He argues that the UK’s scepticism is partly an exposure gap. “In the UK, 99.99% of the population hasn’t reached this technology.”
Getting to the point where they have involves a lot of back and forth, Khastgir admits. “Regulations don’t develop overnight, and society doesn’t have technology overnight,” he says. “It’s a process where engineers, regulators, policy and society need to have a two-way conversation.” He stresses the need for “collaboration to actually bring the technology forward”.
One of the problems is that, in order to advance the tech, and to build up the trust needed to see self-driving vehicles on the road, the engineering sector needs to get better at communicating how AVs work. “Engineers, particularly mechanical engineers, are not great at actually talking about the tech with non-engineers,” he says. And so more work is needed because if engineers can’t translate clearly what a system does and doesn’t do to the public and the people in charge of making the rules, the risk is that self-driving vehicles are stymied before they begin.
Sci-fi made real
The first AVs that hit the road will look an awful lot like the cars and vans we drive today – in large part because they’ll be sat in the same queues as we are and merging into lanes of traditional vehicles. But the longer-term vision is where engineers can really come into their own and start to radically reinvent what a self-driving vehicle might look like.
Dan Sturges, a vehicle designer and author, argues that autonomy is a chance to rethink the transport system, rather than merely automate today’s car. “I myself think we have a huge climate emergency and that we need to advance as fast as possible,” he says.
Sturges worries that autonomous vehicle companies are still trying to cram a new control system into an old form factor. “What they’re basically doing is looking at this profound new way to control vehicles with computers, and they’re trying to fit that into the paradigm of the big automobile,” he says. In his view, the common robotaxi model preferred by the likes of Uber and Waymo in the US, and being tested by Wayve here – of heavy vehicles shuttling single passengers – wastes energy and road space. “They’re just too big,” he explains. “The vehicles are too big.”
The arrival of AVs would also weaken the logic of private ownership, Sturges argues, because if you don’t need to worry about human drivers tiring, then vehicles can be used more intensively in a fleet than they would by households.
If that shift happens, it changes what a vehicle is for. A fleet operator could use mission-specific vehicles that radically change the design of what a car looks like. It could be a two-seat urban runabout that looks a little like a modern car – but perhaps chopped in half – or have an accessible cabin that makes it easier for people in wheelchairs to get about, or a quiet travelling lounge with big screens and TVs for social trips. Sturges says that, at that point, the cabin becomes the product, which makes the interiors, how you get in and out, and how modular and durable they are just as important as performance. “If I was an engineer thinking about this future, they could be engineering mobile living spaces and things that are not cars as we think of them,” he says.
Sturges points to modelling work as a proof of concept for what could happen to our transport network if self-driving cars become the norm. “Yesterday, I was actually looking at a simulation with some engineers at the University of Michigan we’ve been working with, and they’re simulating with a digital twin,” he explains.
“That simulation showed that everyone can go to every place they need to go with only 10% of the vehicles.” The exact number is debatable, and the UK’s travel patterns are not necessarily the same as those in Ann Arbor where the university is based. But the direction of travel is clear: if one vehicle serves many users each day, the fleet can shrink without much impact on our lives.
For engineers, that creates a new set of problems: vehicles that run longer hours need new maintenance regimes. Higher utilisation will increase wear and tear, so components will have to be designed for modular replacement. One simulation suggests that the average robotaxi running around London could drive 170,000km a year, at least 10 times as far as the average vehicle.
In that world, the engineer’s job shifts from building a product that lasts a decade in private ownership to building a system that can survive constant use. And that’s why engineers ought to be excited by, rather than worried about, the AV revolution. “It’s amazing how clean a sheet of paper it presents to us,” says Sturges.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.