The DfT is to carry out a study of driver and road user behaviour to better understand how driverless vehicles can interact with society and other road users.
Claire Perry, transport minister, announced the new study while delivering the keynote presentation at the PACTS autumn conference yesterday (22 Oct) which focused on driverless vehicles.
The new study will run alongside Government trials that will take place next year in three cities, to demonstrate how driverless vehicles can be integrated into everyday life
Ms Perry said the three trial projects will be industry led, with a local authority partner, and will last between 18 and 36 months. The Government hopes to announce next month where the trials will be held, and they could start as early as 1 January 2015.
Ms Perry told delegates at the PACTS conference that driverless technology is “the future”, but stressed it is “important to reassure the public that we are careful of the risk, but also recognising the need for progress”.
Ms Perry said: “Today’s vehicles are so technically advanced that there is the real prospect that driverless cars could be on our roads in a relatively short amount of time.
“What makes this so intriguing isn’t just the technical challenge, it’s the cultural challenge. The idea of tech-enabled driving feels a bit weird. We are all so used to being masters –or mistresses –of the road. Invincible. Always right. Even though it’s our shortcomings that lead to most accidents.
“But if and when it is adopted, this evolution has the power to profoundly change our lives. Not just making driving safer and easier, but reducing congestion, making people more productive – and therefore helping boost our economy too.
“Driverless – or even highly automated – cars and vans can deliver improved safety, improved emissions, reduced noise, optimal usage of road capacity and better use of the scarcest commodities of all these days – time and attention."
Ms Perry added that driver/human error is reported to account for more than 90% of traffic incidents, so it is clear that “driverless cars will make a huge difference”.
She also said there are “great social benefits”, adding: “The advantages of driver–assisting technology for disabled people or those with poor eyesight are clear. I saw a Google video showing a man who was reported to have lost 95% of his vision driving a Google-car.”
Ms Perry also spoke of a future where “driverless buses provide better and more frequent services” by eliminating the cost of the driver. She suggested that a “truly driverless bus could transform rural public transport” and said that one of the country’s major bus companies is “already interested in driverless buses”.
Click here to read the full transcript of Ms Perry’s speech at the PACTS conference.
A driver would have to travel 25,000 miles a day to be exposed to the same risk of death by avoidable accident in hospital in a day. All the “qualification and re-qualification; supervision and management” don’t seem to be working very well so my comparison is more than valid.
Duncan MacKillop, Stratford on Avon
0
Everyone working in the NHS is subject to initial recruitment – a filter mechanism; qualification and re-qualification; supervision and management. Most individual drivers are subject to none of these control measures beyond the entry point filter of passing a driving test.
The example isn’t a valid comparison.
Honor Byford, Chair, Road Safety GB
0
My old advertising guru always recommended that I should change the subject to expose the sentiment so I have taken the liberty of modifying Hugh’s reply. I have changed the subject from road accidents to unneccessary fatalities in the NHS, but you’ll get the idea.
My contention is that, regardless of the ‘system’ in use in the NHS, it is the individual at the end of the chain, whether Doctor or Nurse, that has ‘failed’ where others wouldn’t, through incompetence, lack of judgment, under the influence, over-confidence, arrogance, tiredness, distaction etc. etc…. the list goes on. Unfortunately, we don’t really have time to wait until the ‘system’ is ‘modified’ by some experts, so in the meantime we have laws and penalties to help deal with some of these failings”.
Makes you think doesn’t it.
Duncan MacKillop, Stratford on Avon
0
Yes Duncan but most drivers don’t crash within the first six months or even the next fifty years – and yet it would seem to be the same ‘system’ for everyone.
My contention is that, regardless of the ‘system’, it is the individual at the end of the chain, whether riding the m/c or driving the car, that has ‘failed’ where others wouldn’t, through incompetence, lack of judgment, under the influence, over-confidence, arrogance, tiredness, distaction etc. etc…. the list goes on. Unfortunately, we don’t really have time to wait until the ‘system’ is ‘modified’ by some experts, so in the meantime we have laws and penalties to help deal with some of these failings.
To get back to the news item, it’s the same thinking behind driverless cars – remove the human failings. I’m not 100% convinced yet about the technical aspects, but you can’t blame the authorities for seeing it as a solution in principle.
Hugh Jones, Cheshire
0
The expert panel is there Hugh because I can’t do everything on my own! If you would like a real world example then look no further than the fact that the most likely time for a driver to crash is within the six months following their driving test. The standard behaviourist view would be to blame the driver for one reason or another, but the systems view would ask if the training they received prior to their test was actually doing the job that it was designed to do?
To work out whether or not the training was complicit you look at all the crashes, bumps, scrapes and mear misses that have happened to people within that timeframe. You are looking for all the things about the accidents that were common and all the things that were not common so you could build up a picture of the actual failure modes of the various accident types. Once you had determined the failure modes you would go back to their original training to see if those failure modes had actually been addressed. When you find that they have not been addressed you can then work out ways of ensuring those failure modes are adequately covered in the training and testing process. If on the other hand you find that the training was adequate then you can then look at the design of the roadspace to determine those failure modes and feed that knowledge back to the road designers. This is how it works in every other safety-critical process where you take knowledge from the accidents, work out what the failure modes are, work out how to minimise those failure modes and then feed that knowledge back into the process. You keep doing that until the problem is eliminated, it’s quite simple really.
Duncan MacKillop, Stratford on Avon
0
Ah..more experts – that’ll do it. Sounds as though you’re just passing the buck to yet more quangos Duncan. What will have actually changed in those three years to have brought about your predicted “massive reduction in road accidents and casualties”? Road accidents are ‘real world problems’ caused by real world individuals so at the end of the day, it will still come down to how your experts will address this.
Hugh Jones, Cheshire
0
Three years Hugh!
If I was Minister of Transport it would only take me three weeks to start fixing the problems. My first task would be to appoint Professor Sidney Dekker to be in overall charge of all road transport safety. Second task would be to set up a Road Transport Accident Investigation Branch with the same remit as the Air Accident Investigation Branch.
Third task would be to appoint a panel comprised of leading experts in safety, psychology, statistics, education and engineering but from fields outside of road transport. Fourth task would be to cease funding for all government departments and agencies as well as organisations and charities that promote behavioural safety initiatives. It would probably take around three weeks to complete that lot and by the end of three years you would see a massive reduction in road accidents and casualties.
Sadly though Hugh I am never likely to be Minister of Transport so all those desireable changes will never happen. What I could do however is to continue to work on solving real world problems as if those changes are happening because we must never forget that there are human lives at stake, maybe even yours.
Duncan MacKillop, Stratford on Avon
0
None the wiser I’m afraid Duncan.
Putting it another way, if you were the Minister for Transport and you had say, three years to come up with something constructive that would significantly reduce road accidents, what would you do? What aspects of the ‘system’ would you change? No theories please – some practical solutions.
Hugh Jones, Cheshire
0
To answer Nick and Hugh’s questions there are quite a few things to think about before coming up with some practical solutions to the various problems. The first and most important task is to understand that we are dealing with the whole system not just the individuals operating within it. Any failures within the system therefore are as a result of the interactions of all the components within the system including its rules, regulations, methods and procedures, its design, the environment in which it operates, its failure modes, its outputs and its inputs, its customers and its suppliers and so on and so forth.
If a system suffers a failure and somebody dies then everybody that has anything to do with the system is equally responsible for that failure from the highest levels of government on down. This idea of an “equal responsibility” model rather than the behaviourist “fault & blame” model is the first system modification that I would put in place and from that we can then begin to solve the problem.
Duncan MacKillop, Stratford on Avon
0
Duncan:
What ‘system modifications’ would you propose or like to see put in place, that you think would solve the problem?
Hugh Jones, Cheshire
0
Duncan:
It seems a bit odd to suggest that people have to be in one of two camps – either supporters of behaviour modification or supporters of system modification – when I’m sure most people would agree that both have an important role to play in improving road safety and reducing casualties.
Nick Rawlings, editor, Road Safety News
0
This all cuts to the heart of the road safety problem in that it identifies the two schools of thought on the matter. On the one hand we have the “if everybody followed the rules then everybody would be safe” school of thought which supports the idea of driverless cars. On the other we have the “profound knowledge” school which understands that in complex systems with conflicting goals, humans have to adapt rules to suit the circumstances prevailing at the time.
We could identify the first school as the ‘rigid rule’ group and the second as the ‘adaptable rule’ group. The rigid rule group see people as a problem to be solved and support behaviour modification as the way to make people conform. The adaptable rule group see people as a solution to the problem and see system modification as the key to improving safety.
Of the two schools only the second can bring about any significant and early reduction in road casualties.
Duncan MacKillop, Stratford on Avon
0
Technology has allowed us to communicate faster and with greater accuracy; machines can be built quicker; and manage systems that otherwise would be virtually impossible without them – though the non-computerised mechanical ignition systems on i.c. engines has seen good service for decades.
But with driverless vehicles we are attempting to replace the human brain. Play a computer at chess, and the computer will invariably win. Set it in an environment which demands a learning curve involving all manner of natural obstacles that the human brain learns throughout life to assess and avoid, and I foresee an expensive fail.
What kind of congestion problems might be solved or created by such driverless vehicles? Only a mass experiment on a scale model basis could show us any perceptible results. But even that might be fallible, as the behaviour of the random acts of the organic species could not successfully be incorporated. This amounts to us as a species attempting to play God with a problem solvable through education. Zero KSI on the roads is a goal equal to stopping earthquakes, or trees blown down by gales – it’s not natural. Are we attempting to convert ourselves into machines?
Derek Reynolds, Salop.
0
I agree up to a point Duncan. I am concerned how a driverless car may cope or deal with certain situations as we could but, on the plus side, ‘it’ won’t get drunk, be distracted, lose concentration, get impatient, suffer road rage, want to race someone, speed, tailgate, park illegally etc. all these human failings behind the wheel should be eradicated, which is what Ms Perry is getting at when she refers to “..our shortcomings which lead to most accidents”.
I wonder the same as Paul Biggs – is it the intention that one day we will have no choice but to move around in driverless cars – whether we want to or not? Maybe those who don’t want to, may have to prove that they are 100% competent and safe driving a conventional car.
Hugh Jones, Cheshire
0
It’s not clear from this statement if it will eventually be compulsory for all new cars to be driverless – hence driving would be banned. Such a fundamental change should require consent via a referendum – rather than being presented here as a nonchalant ‘fait accompli’ by a politician. Her example of a man who lost 95% of his vision is bizarre – why doesn’t he just get a taxi? Having the option to push a button so that the car drives itself is one thing – but removing the option to drive at all is quite another.
Paul Biggs, Staffodshire
0
Of course the engine management system does not even require that level of sophistication as the parameters of combustion etc are well known and utterly predictable. Computers are good at that sort of thing, but worse than useless at anything that requires even the most basic reasoning functions.
What a computer program lacks in comparison with human brain function is any form of adaptability in their processing rules. When we drive we are constantly adapting basic rules to suit the prevailing conditions, but this process is so deep in our subconcious that we are hardly aware of it. Without the ability to adapt rules, a situation will quickly arise in a driverless (brainless) car where there is no pre-programmed rule available that exactly suits the current situation. Even the most accident-prone or irresponsible driver can manage this sort of task with ease, but the computer would just freeze with the inevitable and catastrophic results.
Duncan MacKillop, Stratford on Avon
0
Duncan:
A driverless car should be better than some humans are at driving – the reckless and the carefree – but not neccesarily any ‘better’ than those already doing it responsibly, I agree.
Incidentally, would you like to have a go at monitoring the many aspects of your car’s engine management system and make adjustments constantly whilst you’re driving, or are you content to leave it to a computer with the brain power of a fruit-fly?
Hugh Jones, Cheshire
0
Sorry Hugh, but that’s not the case at all!
Unless you are a brain surgeon or a fighter pilot in a dogfight, driving is the most complex spatial task that a human being will ever do. The common understanding is that a driverless car will be ‘better’ than one driven by a human therefore the computer brain behind it must be better than the human one. Even the most advanced artificial intelligence system has a matrix of a couple of thousand neurons and a hundred thousand synapses which doesn’t compare at all well with the human’s 100 billion neurons and several trillions of synapses. As far as developments in computing go we can barely simulate the brain function of a fruit fly let alone anything as complex as the human brain.
By all means put the ‘accident-prone’ into driverless cars if you like, but would you really fancy their chances with a fruit fly doing the driving?
Duncan MacKillop, Stratford on Avon
0
Following on from Duncan’s comment, some people’s mental processing and intellect make them good (i.e. safe, responsible, accident-free) drivers whilst in others it leads to poor (i.e unsafe, irresponsible, accident prone) drivers. Would it not be logical therefore for this latter group of drivers to be the ones in driverless cars?
Hugh Jones, Cheshire
0
Always smile when I read statements such as “driver/human error is reported to account for more than 90% of traffic incidents”. What people that make these sort of statements fail to realise is that the mental processes that result in errors and accidents are the same ones that more often than not result in safety.
The idea that the driverless car will be ‘safer’ because it does not make errors is invalid because a driverless car will have to make errors just like we do simply because it will be forced into making them.
A full explanation of this can be found in the most excellent paper by Thierry Fraichard at: https://hal.inria.fr/file/index/docid/968002/filename/14-rr-fraichard.pdf
Duncan MacKillop, Stratford on Avon
0
Claire Perry confuses ‘drivers’ and ‘passengers’. Driverless cars wouldn’t have drivers would they? All occupants would be passengers.
Paul Biggs, Staffordshire
0
As someone who cares about safety on the road and the devastatiion caused by road accidents and who spent many years trying to do something about it, I should welcome driverless vehicle technology but…the thing is.. I love driving my car!
Hugh Jones, Cheshire
0