Will self-driving vehicles completely eliminate road fatalities?

Humans are the weakest link when it comes to driving. The majority of crashes are due to human error such as inattention and misjudgment; very few crashes are caused by mechanical failure. Self-driving vehicles, then, could be the panacea that road safety has been searching. Except it won’t eliminate deaths on the road.

Motorcyclists

BMW has created an autonomous motorbike which was shown riding by itself in 2018. However, it’s unlikely we will see mainstream adoption of autonomous motorcycles, mopeds and scooters given that the rider shifting their body weight is the way motorbikes are steered. Part of the appeal of motorbike riding is the visceral experience, as opposed to being cocooned in comfort in a car. While many motorbikes are simply commuter bikes, those are at the lower end of the price scale; any autonomous functionality will be well out of that price range for years, especially seeing as anti-lock brakes are only now becoming mainstream.

Motorcyclists are vulnerable due to factors that would be difficult for autonomous vehicles to adapt to:

• Tar bleed in hot or cold weather
• Black ice
• Deep water (especially puddles that hide potholes)
• Sudden crosswinds

It’s likely that motorcycles will get forward collision warning, but less likely they will adopt full autonomous emergency braking or swerving to avoid a collision as there would be too much risk the rider or a pillion passenger would be thrown from the vehicle.

Cyclists, pedestrians and other road users

There are many other road users: electric scooters, cycles, walkers and runners, rollerbladers, skateboarders and more. When they crash on the road, they are included in the statistics. We would need to have a complete separation of their environment from cars and other vehicles, which is not practical.

Then there are specialist vehicles such as mobile cranes and agricultural machinery (combine harvesters, tractors, etc) which have such low production numbers that it’s unlikely there would be enough demand to create autonomous versions immediately.

Historic vehicles

Classic cars, lorries and motorbikes are hobbies for many people and a necessary form of transport for many more. The average age of a car is around 8 years in the UK, so even if every new car that was sold was fully autonomous today, it would take perhaps 16-20 years before the majority of vehicles were fully autonomous. The average age of a lorry is slightly newer, but still over 7 years.

As more and more vehicles became autonomous we should see a reduction in crashes and in the severity of crashes, but it would take considerable time.

Mechanical failure

A small proportion of accidents are due to mechanical failure, some avoidable, some unavoidable. These include things like brake failure, punctures, the chain snapping (on a motorbike), a heavy trailer detaching from a tractor unit or a wheel-off incident (this can cause a loss of control, or the wheel itself can become a missile which causes a fatality).

Unavoidable accidents

A small proportion of accidents will always be unavoidable: a landslide pushes you off the road, a pedestrian runs out in front of you, etc. Autonomous features may be able to reduce the chance, but they don’t eliminate it.

Avoidable accidents

Drivers have a responsibility to ensure their vehicle is roadworthy and that any trailer and load are secure. This doesn’t always happen, though, and avoidable crashes can occur when a trailer detaches or a load falls from a vehicle.

Cars and lorries

That leaves cars and lorries. There’s a real impetus for autonomous lorries to be developed due to the chronic shortage of drivers. Autonomous car technology is becoming more affordable and being included on an increasing number of vehicles at various levels. We may be at least 10 years away from every new vehicle having full autonomous driving features, though, as the technology begins to struggle once the road surface degrades and weather/light conditions deteriorate. It’s likely that, at least for the next couple of decades, even fully autonomous vehicles will have to be driven at some times.

Therefore, in the short-term, autonomous vehicle technology will begin to make an impact, reducing the severity of crashes and saving some lives. In the medium-term, regions with high adoption of autonomous vehicles that separate vehicle and pedestrian traffic should see zero fatalities. In the long-term, widespread adoption of self-driving vehicles will contribute to a trend towards zero with road fatalities, but stubborn areas will remain and there will be high-risk individuals who opt out of the autonomous features.