I read an excellent report recently from the TRL Academy - Transport 2020: Addressing future mobility needs which sets out expert views on many of the challenges for society as we move towards more autonomous and connected vehicles. The report poses lots of questions – for many it will take several years before we have a definitive answer.
The challenges are multiple and varied – will communication networks be able to handle the Terabytes of data that an autonomous car will generate ever second?
Reliable connectivity is essential for effective management of automated vehicles, but the experts point out that no vehicle should depend on connectivity for its safe operation. As in other sectors, highly or fully automated vehicles must always have a safe fall-back level of operation, so that the vehicle can safely protect its occupants – but will those occupants suddenly have to be involved just because the mobile phone signal stopped or the satellite link goes down?
The prize of eliminating human error from road transport is of course one worth pursuing but even the experts agree that automation can never completely do this.
However, it offers the potential to tackle key risk factors on our roads such as driver fatigue, impairment and inattention. At IAM RoadSmart we are obviously concerned to know if we have a future coaching humans, but the TRL Academy experts do agree that there is a risk of drivers becoming de-skilled, or relaxed to the point where they are unable to take over at short notice. This is where we may still be able to be of service!
Some of the biggest moral dilemmas that still need to be resolved are linked to the fact that artificial intelligence (AI) will be essential for the introduction of autonomous vehicles, because “driving is a known, but unpredictable, environment owing to changes in weather; the behaviour of the driver and vehicle occupants and the behaviour of other road users.”
At least it does sound like these folks have been on the M25 at 5.30 on a Friday evening where ‘unpredictable’ can be an understatement. Will AI ever replace that experiential behaviour we all take for granted such as waving a car out from a side road or just ‘knowing’ what that driver is about to do. Will human drivers bully autonomous cars – the evidence so far is limited but mixed.
It could be that as soon as drivers know a driverless car will stop if you approach it then some will turn that to their advantage in queues and at junctions. The transition phase when human controlled and driverless cars share the roads is going to be very illuminating and likey to last many years.
And then there is the ultimate moral question for any autonomous car user – do you want to be in a vehicle that choses for you who to hit in an inevitable crash situation. Head on with the truck or swerve into the cyclist?
Autonomous cars offer the potential to increase productivity by freeing ‘drivers’ to focus on work-related tasks, child care, or social engagement. They could also offer much greater mobility to people who are currently unable to drive.
Some recent US research however suggests that most drivers will still want to watch what the car is doing on their behalf rather than engage in other tasks – it will take time to build trust. Society will also have to answer questions such as how old can a driverless car occupant be – will we send the kids off on the school run by themselves?
One of the most interesting statements in the report is that “there is currently a strong technology push for autonomous vehicles rather than a societal pull,” are we as consumers actually asking for autonomous vehicles or are they coming because suppliers want to sell us new tech?
If we do give up control then the answers to all of these questions will have to be resolved. At IAM RoadSmart we are running an expert conference in October to try and provide some of the solutions.
By Neil Greig, IAM RoadSmart director of policy and research