March 2017

Singapore is gradually introducing driverless cars and taxis on public roads. Are they the answer to road safety?

1st Place - Ng Chee Siang

Yes, driverless cars are the answer to road safety, but they are only one-half of the answer. The drivers and the pedestrians are the other half of the answer to the problem of road safety.

 

 

Cars have been getting safer since they have been pushed into existence. From regulatory changes requiring every car to have seat belts and every driver to pass a driving course to air bags and structurally stronger cars, much has been done to ensure the safety of the driver and the passengers inside. Traffic lights, zebra crossings, overhead bridges and underground tunnels have been built to protect pedestrians from vehicles on the road. However, they have not stopped accidents from decreasing. In fact, statistics have shown that traffic accident rates have gone up since seatbelts became compulsory, however, the fatality rate for each accident has gone down (Freakonomics). The issue boils down to reckless drivers and pedestrians. Drivers drive more reckless because they feel that their cars can protect their lives better while pedestrians feel that these traffic lights and crossing acts as a shield against incoming vehicles.

 

 

Driverless vehicles are just part of the routine upgrade towards making the car safer for the drivers and passengers, except this also makes the car safer for every user on the road. The main cause of traffic accidents is human-factor – inside the car or outside the car. Driverless cars remove the human-factor inside the car, preventing drunk and/or reckless driving from occurring. While it is unable to prevent the human-factor outside the car (eg; someone jaywalking on the road), it can detect and react faster and better than a human can, thus able to provide some form of protection for both the driver and the jaywalker.

 

 

Although driverless cars can significantly reduce the traffic accident rates, it needs to be widely accepted before it can become a real safety measure. If a road is totally filled with driverless vehicles, it is probably 99% safe – with the 1% caused by jaywalkers. However, if the roads are filled with some driverless and some with-driver cars, it might not be any much safer than having the roads all filled with with-driver cars. Two driverless vehicles might be able to avoid the collision because the two probably figured out a path for each other and communicated with one another to ensure they do not clash in those paths. One driverless and one with-driver car could end up in a collision because they do not know which path the other party is taking. While many agree that a car should sacrifice the people in the car if it can save more lives on the road, they have also said that they would not buy such a car (The Guardian, 2016).

 

 

There is also the issue of hacking when it comes to driverless vehicles. These vehicles could be hacked and their controls could be overwritten by viruses, result in them being a big hazard when on the roads. Current cars are hack-proof because they are non-tech products. Driverless cars are heavily dependent on traffic data and have software in them to operate optimally and safely. However, this software is updated through the internet, which makes them susceptible to hacking. 

 

 

While it may not increase road safety by much, driverless cars would increase the economic activity of Singapore and the world. It is increasingly getting harder to fill jobs in the transport industry. The youths of today do not wish to be working in these blue-collar jobs – bus or cab drivers, or train operators. The way forward would be to automate away these jobs so that we can free up human resources to work on things they are more interested in doing. Driverless vehicles are also a way of keeping cost down. In the transport industry, manpower is one of the top operating cost, and the pay has to keep going up in order to attract and retain drivers. However, the high cost is also passed on to consumers, who in turn have to pay a higher price for these transport services. Driverless vehicles remove the need for a driver or operator, thus reducing the cost of transportation.

 

 

There are many answers to making our roads safer and driverless vehicle is one of them. But it is not the full answer. The public – drivers and pedestrians, needs to be careful when it comes to the road. Driverless cars can only reduce traffic accidents to a certain extent. If everyone starts jaywalker, driverless cars will still have to sacrifice either the driver or the pedestrians. However, driverless cars serve more of an economic benefit than a safety benefit. Driverless cars can free up human resources to perform other job and lower transport costs.

2nd Place - Joshua Goh

The idea of driverless vehicles is nothing new to the layman, while we have driverless trains present on the Downtown and Circle MRT Lines, the demands on driverless cars are much more complex. Unlike trains, cars are not limited to a set of rails but are in a dynamic environment where they must interact with other vehicles, change lanes and avoid oncoming traffic when turning. Therefore, the technology behind driverless cars is infinitely more complicated than what we having running on underground rails in Singapore.

 

However, such technological complexity has been argued as a necessary advancement in road safety. The allure of autonomous vehicles is that they can, by wireless signals and sensors, better communicate with other vehicles to allow for more efficient traffic flow. In 2014, researchers from the University of Exeter found that a seemingly minor event on the road, such as a vehicle suddenly changing lane would have a “butterfly effect” in causing the human driver behind it to slow then and the vehicles even further behind to slow down even more. This explained the occurrence of large jams that would suddenly occur with no obvious reason for their cause. This created a ‘backward travelling wave’ where many human drivers slowed down in a chain reaction to the vehicle in front of them, thus continuing the wave and the jam. This is also due to the limited ability of human drivers to communicate with other human drivers beyond signal lights and angry honking.

 

A community of purely driverless cars should be able to solve this problem with information being shared among a network of driverless vehicles to prevent such unnecessary traffic deceleration. Therefore, driverless vehicles, in being able to anticipate the movements of other vehicles with sensors and being unrestrained by human reaction time, should logically ensure better safety on the roads. Nonetheless there are two issues with this scenario, (1) current technology is not developed enough to offer driverless vehicles on such a large scale to dramatically increase safety and (2) the co-existence of human drivers and driverless cars on the same road still poses risks to road safety.

 

Imperfect Technology 

 

A driverless car is not fully autonomous, SAE International has come up with a chart to denote the spectrum of autonomous levels of driverless cars, ranging from Level 0 with “no automation” to Level 6 “full automation” and no human input. Therefore, in between, a certain degree of human exists across the remaining levels. This state would probably be the norm for the short-term given the enormous among of data and programming that is required for a fully-autonomous vehicle. This means that for most driverless cars, a person is still required to be present as a safety feature. However, this does not necessarily enhance safety. In a Bloomberg feature on Ford’s development of driverless cars, Ford noted that "high levels of automation without full autonomy capability could provide a false sense of security", therefore a backup driver might find it difficult to react to a situation after sitting idle in the driver’s seat for so long.

 

The solution therefore appears to be to make all driverless vehicles Level 6. Nonetheless this will not automatically make the roads safer since arguably the roads are as safe as the least safe driver on them. This is the human factor that cannot be driven away from the issue of autonomous vehicles. Unless there is a complete conversion of all vehicles to driverless modes, humans and robots would share the same roads and the risks present in driving.

 

Interacting with human drivers


Envisioning a future with completely-autonomous cars plying the roads seems like an absolute guarantee for road safety. However we are hardly anywhere near this state and it is likely that for the short-term at least, driverless and driver-operated cars will have to share the same roads. This mix has created a situation where accidents are still able to occur.


In August 2016, nuTonomy began road trials of what was touted as the world's first on-demand driverless taxis. However in October 2016, the self-driving taxi was involved in a minor accident with a lorry. Responding to the accident, Dr Park Byung Joon stated that such accidents are to be expected as "Humans don't always behave the way they should on the roads. And technology is not advanced enough to pre-empt how humans would behave."

 

Conclusion

 

Therefore driverless cars in their current state are unlikely to improve road safety in the short-term as they have to co-exist with human drivers and contend with their own technological development. Dr Park downplays Singapore’s thinking that autonomous vehicles will soon be a common sight, giving it at least a decade more before they become regular features on Singapore’s roads. Thus we should not see driverless vehicles as the eventual solution to road safety, instead we should continue investing and developing road safety habits among current road users as it is better to lose such a skill when driverless vehicles become a common reality than to lose it in anticipation of them.