John Yellowlees: There’s a long way to go before robots take over the wheel

London’s Docklands Light Railway has offered panoramic views unhindered by any cab since 1987, Glasgow Subway’s new trains will be without onboard staff, and buses that drive themselves will soon take commuters over the Forth Road Bridge between Edinburgh Park and Ferry Toll.

These public transport applications abide by the most stringent safety standards, since people who accept the risk inherent in driving themselves are much more demanding when someone – or something – else is driving. So it will be as autonomous driving is rolled out to private cars.

Even when the technology reaches an acceptable standard, showing that AVs are safe will entail massive testing on the road. Driverless vehicles will have to negotiate local road networks at either end of their journeys, so it is insufficient to test them only on motorways or in places like business parks. Using an app to hail a robot-driven taxi would work only if they could cope on cramped street layouts, with drivers coming in the other direction.

Wisely, the Centre for Connected and Autonomous Vehicles asked the Law Commissions of England, Wales and Scotland to undertake a far-reaching review of the UK’s legal framework for automated vehicles and their use in public transport and on-demand passenger services.

Firstly, the Commissions considered how to make vehicles safe by design, through the establishment of a safety assurance scheme to complement the current system of international type-approval.

A second consultation has proposed making vehicles safe by operation, through focusing on the role of the operator to ensure that fleets are managed and supervised appropriately in the absence of a user-in-charge, with a new single national system of operator licensing, placing operators under a legal obligation to ensure roadworthiness.

The Commissions do not propose to regulate fares for hire of AVs – instead, they think that consumers should have the opportunity to compare prices before booking.

Having an AV for exclusive use does not necessarily require the consumer to invest capital in buying the vehicle outright, a move that carries financial risks, so initially they anticipate that consumers may enter into a leasing arrangement, with the leasing company being the registered keeper and therefore responsible for roadworthiness.

There are potential economic advantages from driverless operation – autonomous vehicles come to you and there is no driver to pay, so the Commissions see them as combining the economy of car clubs with the convenience of taxis.

Freed from the necessity of car ownership with its large sunk costs, people may think differently about transport options, making greater use of public transport and active travel.

AVs could thus reduce congestion, especially if rides are shared. Their much better utilisation (the average car is parked 96 per cent of the time) could declutter towns by reclaiming land used for parking. Hours presently spent behind the wheel could be used more productively.

A transport system that works better for disabled and older people works better for all, and the introduction of AVs could help give them the same access to transport as everyone else. Without the fixed costs of a driver, it would become economic to run a fleet of smaller buses autonomously at more frequent intervals.

There remain many unresolved concerns about AVs, including a risk that they might freeze when confronted with unexpected weather conditions or with unknown objects such as plastic bags or even leaves.

Some people with disabilities may miss assistance provided by the taxi driver in getting in and out of the vehicle. Developers planning remote supervision will require connectivity and suitably trained staff. Rural roads will be particularly difficult with their poorer surfaces, sometimes with passing places.

There will be new maintenance challenges to ensure that software is kept updated and cybersecurity protected – it is not yet clear whether the technology will be sufficiently safe for individuals to be able to organise supervision, software updates and security for themselves, and there may be considerable ongoing costs for updates, repairs and servicing.

Against their better utilisation must be offset the additional traffic represented by vehicles returning empty after dropping someone off or running round the block while the customer is shopping.

The driverless cars of the future will be electric and will offer an environment sympathetic to their occupants’ mood, but they probably won’t go any more quickly than at present – and route-setting while under the influence of alcohol ought to be no more acceptable than drink-driving is now.

The public will want to see real benefits before it is willing to pay extra costs. So the future may not be so different after all.

John Yellowlees, Scottish chair CILT.