driverless cars and taxis - will it work?

why driverless cars and taxis will never make an impact on british roads

Working in the taxi and minicab industry I see a lot of change going on, and currently there are a lot of things that look almost overwhelming. The arrival of a driverless car, for example, is something that would send a shiver down most minicab and black cab drivers' backs. With companies like Apple, Google and Uber all working on different driverless car technology, you'd be forgiven for thinking the end is near for taxi drivers' livelihoods. However, if you take a step back for a moment, you can quickly see that the possibility for driverless cars and taxis to be fully functional on Britain's roads is very limited, and at best, a long way off.

Driving on the streets of the capital requires more than just a brain, and is not for the faint hearted. You need aggression and common sense; you need to use your wits and think on your pedals; you often need firm eye contact, if you are to get out of a busy junction for example. There is a limit to how much aggression you can program into a car before it just becomes dangerous. Picture it - some poor passenger sat in a driverless car at rush hour, trapped between a box junction, a green light, a van, 100 pedestrians and 10 cyclists. You're going nowhere fast, but with this driverless "solution" you've introduced a new problem. The driverless cars quickly become mobile obstacles. Unable to force their way into traffic, they would end up holding up other cars and bring a capital city to complete gridlock. I'm sure a Google engineer reading this would be thinking: "No! We can handle that. I can prove you wrong." And they probably can in a controlled environment, but this doesn't take into consideration one major factor - the competition. Every black cab driver and minicab driver in the capital would make life hell for these robots. When driving in the capital you need other drivers to give you a little help now and again, but if everybody around you hated you, you would be unable to move. It does not matter how good the technology is, this is one obstacle that will never be overcome.

Accidents do happen

With almost 100 accidents and collisions on London's road every day, it's highly likely that these driverless cars are going to have an accident. Let's say a Google car pulls out from a junction and its radar fails to spot a cyclist come past traffic at around 30mph. The cyclist takes a nasty tumble and breaks a few bones. Who's at fault and who is going to pay out? The first question is why the Google car pulled out in the first place. Was it a result of the programmed aggression within the engine that the car needs to navigate through London roads? Or did the radar simply not pick it up due to an internal fault? Was the view of cyclist obscured from the car's camera by a stationary vehicle? Google's driverless car accident rate is 0.65 accidents per 100,000 miles compared with 0.3 accidents in cars driven by people. When a new aeroplane has an unexplained accident, it's not uncommon for the whole fleet to be grounded while investigators determine the cause. In the interests of public safety, you could not have the driverless car on the road while the accident is investigated. How many accidents will they be allowed to have before they are deemed unsafe for ongoing use?

You can't avoid congestion

Every capital city in the world is over congested. The number of cars on the road increases year-on-year, and with each passing month, driving around London feels like it's getting a little more difficult. Boris Johnson has already suggested a cap on the number of Uber drivers allowed on the roads to try and limit numbers, and it's a logical step for everyone to start looking at capping the number of vehicles allowed on the roads at any one point in time. Congestion charging is a part-solution, but you can't just continue to increase the charge until only the rich are able to afford it. What's needed is a simple system which is fair for all, and that does not depend on how deep your pocket is. Driverless minicabs for London's roads will simply turn up too late to the party - you can't just keep adding vehicles to the streets. In Uber's case, it's not hard to imagine that they would simply replace their existing drivers with the driverless cars, but I can't see that working; that driver still has a license and Uber doesn't own them, meaning that simply saying "we won't use that driver anymore" does not remove them from the roads. The government could introduce some form of public transport license for driverless cars, and claim that the tax raised will go towards improving public transport - but how would that work if vehicle caps have already been introduced?

The taxi and minicab industry is very unique, and too many people would be affected negatively by fleets of driverless taxis. Taxi drivers are almost universally self-employed, free to work for who they want and when they want within reason. Every taxi and minicab in town would become a moving obstacle for a driverless car. Anybody using one would quickly get the impression that they have become public enemy number one. If you can't provide a service then you don't have any customers, and your business model quickly falls down.

Getting up to speed

Speed is also the enemy of a driverless car - they just can't go the distance. The faster they go, the harder they have to work at scanning for obstacles and identifying potential dangers. To combat this, their speed would need to be carefully regulated. But not only that; even if speed was capped, how safe would you feel to be on a motorway doing 70mph without a driver? At this sort of speed the, chance of death from an accident becomes much greater. And of course with speed comes distance, which, due to their battery powered engines, is the driverless car's next worst enemy. Battery powered cars are getting better with technology, but a battery powered car driven by a person is like a mobile phone on standby, whereas a battery powered car driven by a robot is like a mobile phone playing a Youtube video; battery power use would go through the roof, due to the amount of power hungry devices working to keep the car moving. Batteries and CPU power are both obstacles that can ultimately be overcome though.

Things change

Roadworks, closed roads, RTC's and hand signals: driverless cars need to understand and be responsive to the environment they are in. Small and temporary changes to road infrastructure represent a big challenge. A police officer attending a traffic accident will often use hand signals to direct traffic, but anybody who has used an Xbox Kinect will know that hand signals to a computer are not always easily interpreted outside of a calibrated and controlled environment. How well will a driverless car pick up these signals at a distance on a dark and rainy day, and how will the computer even identify the person as a police officer? Or what about a traffic junction when the traffic lights are not working, which is potentially a far more dangerous and hazardous situation. And what about a plastic bag? Plastic bags look no different from a large rock to a computer sensor, and the vehicle would likely be forced to take avoiding action.

The British weather

Put simply, there has not been enough testing of the cars in bad weather. The cars are being used and tested in California, which is currently going through one of the most severe droughts on record, so it's unlikely the vehicles have ever even seen rain or fog, especially not the likes we are used to in the UK. A heavy snow storm could keep driverless cars off the roads for weeks. You can't provide a commercially viable taxi service if you can't guarantee it will operate.

Don't believe the rhetoric

Whenever I see driverless cars pitched, I see them pushing the potential upsides; less accidents, less congestion and less pollution. Less accidents is possible, but less congestion? Absolutely not. You don't reduce congestion by adding more cars on the road, and the truth is that with the assistance of other road users they will just become slow moving obstacles. Less pollution? Just because something is electric does not make it environmentally friendly. We still have to burn coal, oil and gas to generate the electric. Also, driverless cars aren't the only electric vehicles, and the rising prevalence of electric cars has nothing to do with driverless technology.

I can understand why people are on the driverless car bandwagon - if you work in the industry you need to be talking it up, selling the idea and telling people it will work, but I would bet that these people also go home at night wondering if this is all really going to be possible. Driverless technology at best is a great aid to normal cars driven by real people, but will they ever become part of a mainstream taxi service? I don't think so. They may play a part in small scale ground transport in more easily controlled environments, but when it comes to inner city roads and motorways, it seems unlikely.

Technology for driverless cars over the next five to ten years will mostly be limited to driver aids which still allow a human to take over control. We may well see a small scale role out of driverless taxi services in more controlled environments, such as from an airport car park to a terminal or in an adaptation of bus routes, wherein the environment can be more tightly controlled, but I don't think we will see the tech start to make major inroads in the taxi industry for at least around 15 years. Testing in major cities will raise a number of issues which will delay the commercial rollout of the technology, but it's clear that driverless cars will struggle to make a real impact on our roads, at least in the foreseeable future.

About the Author

Jonathan Kettle is a serial entrepreneur and the founder of successful online taxi-booking business He is also the founder of and a co-director of Web3 Resource Ltd.