In the UK in 2013, there were 1713 road deaths. To put this into perspective, in the UK 6000 people a year die from accidents in the home, around the same number commit suicide, there are around 8700 deaths related to alcohol abuse, and about 150 people are killed at work. This is not to minimize the issue of road deaths as it does represent one of the leading causes of accidental death.
Motorcyclists are a high risk group simply because they have comparatively little protection. They account for 1% of road traffic but 20% of all casualties. About half of motorcyclist deaths are due to other vehicles failing to see them at junctions, but it also an unfortunate fact that the other 50% are largely or wholly down to the rider. At least 25% of motorcycle deaths involve no other vehicle, typically when a rider simply fails to get round a bend. Other major causes are filtering through traffic, lack of judgement when overtaking and collision with road furniture. The phenomena of ‘motion camouflage’ is thought to be one cause for the number of accidents at junctions, something that few drivers are aware of.
According to figures from the ONS in the UK, an average of 2-3 pedestrians are killed each year by collision with cyclists. From data published by Transport for London, in 2007 10% of pedestrian deaths in London were caused by ‘powered two wheelers’, 3% by pedal cycles, 12% by heavy and medium goods vehicles and 16% by coaches and buses. Cars and light goods vehicles account for 55%. Total pedestrian deaths in the UK run at around 23% of all road casualties. If we assume that the data from London is typical of the country as a whole then 45% of pedestrian casualties are not caused by cars.
According to ROSPA, 16% of fatal or serious cyclist accidents reported to the police do not involve a collision with another vehicle, but are caused by the rider losing control of their bicycle. In collisions with other vehicles, the blame probably splits 60/ 40, where 40% are due to the cyclist not looking properly. Now it could be argued that autonomous systems might still be able to save lives in these cases by reacting quicker or better than a human driver. This is of course pure speculation, but for the purposes of this note assume that about half could be prevented by this means, leaving a residual of around 20% (of 84%). In 2013 109 pedal cyclists were killed.
HGV’s account for around 20% of deaths on ‘A’ roads but 50% of deaths on motorways. In 2013 100 people were killed on UK motorways. We can therefore ascribe 50 deaths to HGV’s on motorways. With total car driver/passenger deaths of 785, this gives an estimate of 137 (20% of (785 – 100)) for non-motorway car/car passenger deaths, and a total due to HGV’s of 187.
It is absolutely clear that even if changing the entire car population to fully autonomous vehicles resulted in zero car accidents this would still not eliminate road deaths. Based on the figure of 1713 deaths in 2013, we would still have something like:
177 pedestrian deaths
165 motorcyclist deaths
36 cyclist deaths
187 car occupants killed by HGV’s
This gives a total of 565 people killed on UK roads that are not due to cars. That leaves an estimated 1148 deaths that are due to car driving, and that it might be possible to eliminate through improvements in car design. The design of road systems has a major influence on accident rates though it is unclear what proportion of road deaths better design could eliminate.
In the UK there are some 31.5 million cars, and the average annual mileage is reported to be around 8000 miles. Taking a figure of 1148 deaths per year, this equates to 1 death per 220 million miles that we can attribute to car driving.
Of accidents involving cars in the UK:
20% of fatal accidents are sleep related
20% involve a driver over the drink/drive limit
in 38% of deaths the victim was not wearing a seat belt
young drivers are twice as likely as other age groups to have an accident; they are at higher risk because they tend to be over confident, and their risk perception and appreciation of hazards is inadequately developed.
44% of fatal accidents involve a vehicle going out of control at a bend, particularly by young drivers for whom this is probably the leading cause of death
under inflated tyres cause around 30-40 deaths per year.
in 17% of deaths exceeding the speed limit was a factor
based on data from the USA, weather may be factor in around 20% of cases
in 2014, 24 fatalities were linked to the use of mobile phones – surveys indicate that at any time more than 1% of drivers are using hand held mobiles, with the under 30 age group being the worst culprits
These numbers are not simply additive because accidents can involve multiple factors, but they indicate the scale of reductions that might be possible through driver aids.
Technology already exists to tackle some of these problems. Tyre pressure readings are or will be mandatory on all cars. ADAS’s such as lane departure warning, adaptive cruise control, adaptive light control, automatic braking, blind spot detection, driver drowsiness detection, speed limit warnings, night vision aids, all these functions are gradually being introduced into mass market cars and will help to reduce accidents. The detection systems being developed for autonomous cars can also be used to provide warning systems for drivers, for example at junctions and for lane changing. Even simple things such as educating both drivers and motorcyclists on the hazards of motion camouflage would help, perhaps along with measures to improve rider visibility.
To tackle unlawful driving of vehicles ignition interlocks may be able to reduce the incidence of drink driving. Interlocks to enforce the wearing of seat belts do not seem to present an insurmountable problem, though this would be expected to reduce death and injury rather than the number of accidents.
The best estimate is that seatbelts reduce the probability of death by 50%. Enforcing seatbelt wearing might therefore save around 149 of car driver/passenger deaths (0.38 x 0.5 x 785).
If driver aids can reduce the incidence of sleep related accidents by (say) 75% that would save a further 172 deaths. A similar gain on drink driving would also save 172 deaths.
So we might reasonably guess the following reduction:
enforcing seatbelts – 149
limiting drink driving (75%) – 172
limiting sleep related accidents (75%) – 172
eliminating hand held mobile use – 24
reducing motorcycle accidents at junctions (75%) – 124
reducing accidents due to underinflated tyres (75%) – 26
reducing excessive speeding (50%) – 98
total reduction – 765
If we round the total down to make some allowance for there being multiple factors at play in road deaths, then this gives an estimate of at least 700 fewer deaths, along with a corresponding reduction in overall accident and injury rates. This figure does not specifically include any allowance for reductions in accidents due to systems such as lane departure, or automatic braking.
This still leaves around 450 deaths per year due to cars, but it shows that there is no reason to suppose that the wide spread adoption of ADAS’s (existing and those that could be reasonably anticipated) cannot drastically reduce road casualties caused by cars, that the general decline in roads deaths cannot be continued, irrespective of the introduction of semi or fully autonomous cars.
One point to bear in mind when considering the potential of autonomous systems to reduce accidents is to remember that semi-autonomous systems entering the market now are most applicable to motorway driving, which is already very safe.
In the UK 4% of drivers are uninsured and many of these may well also be driving with no MOT and having not paid road fund tax. With the advent of ‘connected’ cars, in principle at least this should be simple to stop. Research shows that uninsured drivers are more likely to be involved in accidents and more likely to be involved in drink driving.
Comments