Self-driving Cars and the Future of Transportation
Chances are that in few years driving one’s car will share the fate of horse riding and become a luxurious hobby rather than a means of transport.
Elon Musk, CEO of Tesla-the company producing electric cars-claims that in the course of our lifetime driving one’s car will become illegal. Next generations will be wondering how could we let humans drive vehicles that weigh few tones, at the speeds that give us just a fraction of a second to react and avoid accidents. This fraction of a second can be enough for computers though; to analyze the current situation, take into consideration various scenarios and then choose the optimal one.The question is not IF but WHEN will they replace humans at the steering wheel.
Letting go of the steering wheel
Working on autonomous cars (or self-driving cars as some people like to call them) is nothing new-it’s been on the agenda since the 1980s but the last few years have brought the biggest progress due to the rapid development of the computer natural image processing. It was not so long ago when computers had troubles recognizing the image on a photography or in a movie. Thanks to the progress in developing artificial intelligence algorithms, now computers can identify objects in a photo faster and better than people.
Apart from the GPS receiver that can easily track car’s location, autonomous cars are equipped with various devices that analyze the surroundings. One of them is the so-called “lidar” - the device that uses laser light to determine the location and distance of the objects around the car. Few video cameras constantly analyze the image, recognize the street signs and identify the roadway markings. Classic sensors assist while parking and activate the brakes when there’s a risk of collision. The data from all the sensors is analyzed by the onboard computer and based on it, the decision determining wheels motion and speed change is made.
When it comes to creating the indispensable solutions for self-driving cars, the most advanced companies include: Waymo - part of the Alphabet holding company (the owner of Google), Mobileye - Israel’s company bought for $15 billion in March 2017 by Intel and Tesla - the leader in electric cars’ market. Apple is also eager to use the market’s potential but their plans remain a secret.
The Society of Automotive Engineers suggested a 6-level scale of car’s driving automation. Level 0 means no automation at all, level 1 - a partial support while driving, e.g: cruise control, sending alerts while missing one’s side of the road or support in parallel parking manoeuvre that can be done by the computer itself. This kind of support is available in the growing number of high-end cars.
Level 2 covers the possibility of the onboard computer taking control over steering the car, its velocity and direction but the driver has to be focused on driving all the time. He should keep his hands on a steering wheel and be ready to intervene and take control of the vehicle at any moment. Level 3 means that in certain circumstances (e.g. certain visibility, weather, and road conditions) the car can drive on its own. Level 4 equals car’s autonomy in the most of the conditions. Level 5 means that passengers simply set the destination but the ride goes on without any human intervention. Such cars won’t even have steering wheels.
Level 2 and 3, which involve the driver constantly monitoring the ride, are extremely risky. Shortly after October 2014, when Tesla offered its clients the possibility to drive with autopilot in certain conditions, you could have found short films on the web with drivers watching movies, playing cards or… taking naps during trips. Autopilot seemed to work so well that drivers deconcentrated and ignored the fact that they were still responsible for controlling the vehicle.
Tragically for 40-year old Joshua Brown, he had too much faith in technology. On the 7th of May 2016, he was driving a freeway in Tesla Model S through the northern Florida. The conditions were good so he activated autopilot mode. The upcoming semitrailer truck began turning left, crossing Tesla’s path. The driver didn’t react, the emergency braking system didn’t “register” the semitrailer and continued driving. The upper part of the car was cut off during the collision at the speed of 119 kilometers per hour, and the bottom part stopped in the field, outside of the road, after battering two fences. Brown died on the spot. The detailed investigation proved that the onboard computer mixed the white semitrailer with the light sky behind. We don’t know why the driver didn’t react at all. Maybe he was watching DVD movies at that time? There was a DVD player found in the car wreckage. In Tesla’s official statement one can find all the data proving that the rate of accidents happening in the cars with autopilot mode activated is lower than in those driven by humans.
No traffic jams, no accidents, no fun
Many of the automobile producers decided to start sales after reaching Level 5 - meaning full automation. Such cars may become available on the market in 2021. This could be a revolution comparable to the rise of Internet and mobile phones. What it’s supposed to bring is the reduction of road incidents, their victims, and costs related. According to McKinsey & Co, we can expect even 90% fewer road accidents to happen. In the USA alone this could save thousands of lives and bring $190 billion in savings a year. In 2016 in Poland, police noted 33 thousand road accidents, claiming 3 thousand victims and over 40 thousand injured.
What will also disappear? Traffic jams. Freed from human brains’ limitations (like reaction time) cars will go faster and start at the same moment. Also - road traffic will be smoother. After some time even traffic lights will be gone and cars will be communicating with each other to decide on the mode of driving. Self-driving cars will be used by the elderly, the disabled, kids and those temporarily unable to drive, e.g. during Friday’s night. Drivers will be able to reclaim the time wasted on driving and will have the chance to spend it playing and having fun.
And what about those who like driving? Well, it is very probable that some new theme parks will emerge offering sitting behind the wheel of the traditional car for a certain price and after signing relevant waivers and agreeing to wear helmets and pads.
Transportations always influenced changes in the cities. At first, towns have emerged along rivers. After the railway came, cities reached by the trains developed the most. The appearance of cars allowed people to move out to the outskirts of the city. As a result, the industrial areas became separated from the residential ones.
Yet another major change will be brought by self-driving cars. Never-ending desperate search for a parking place will be gone. Passengers will be able to leave the car after reaching the destination and it will drive to the nearest parking lot on its own (which in this case could be located in the far less convenient place), waiting patiently for its owner to call it back. Parking lots will make room for buildings or urban green.
But if we could call our car anytime we want, do we really need to own one? According to Paul Bartner, transportation advisor, cars are parked for 95% of the time. Instead of buying and maintaining a car, one could make use of the available services; rent a vehicle suitable for the given moment. Summer date? Convertible, please. Family weekend in the countryside? A 7-seater and room for luggage. Service will cover maintaining the fleet. The number of vehicles per person will drop significantly.
Victims of the revolution
The upcoming changes are not eagerly awaited by the transportation companies’ workers. According to PwC report, there are 650 thousand professional drivers in Poland (3,5 mln in the USA). Will their work become redundant? How will they deal with the loss of jobs? And what about car shops suffering from a dropping number of customers? It seems that the rise of autonomous cars will bring a lot of discussions and protests.
Some also claim that lowering the cost of travel will encourage people to move out even further outside the city. This, in turn, may bring bigger road traffic and a greater congestion around the city centers.
It’s still undecided who would take responsibility for the potential damages done by self-driving cars. The legislation is still far behind the technological progress. The passengers would not be held responsible for sure as they have no influence over the course of events. The same goes for the owner, whose only decision concerns the model of the car he chooses. He would have to make sure though that the car is serviced regularly and - what is crucial in this case - that its operating system is up-to-date. Car producer would be blamed only if the accident happened due to some technical faults. As for on-board computer’s wrong decisions, the responsibility would have to held by the producer. If car insurance for individual customers would not be needed anymore, this could also revolutionize the insurance market.
In January 2017, European Parliament Committee on Legal Affairs approved the report ensuring the creation of specific legal framework concerning robotics and driverless cars. What’s interesting, the preamble to this official document mentions Frankenstein and Golem as the symbols of the long-time human efforts to create new beings. The document also covers the obligatory registration of all the robots, the need for a built-in, global power-off button protecting them from wrongdoings and the need for the insurance framework. It seems that most far-reaching idea involves the creation of an “electronic personhood” (along with the natural and legal persons) - a new legal entity. This status could be granted only to the most advanced robots.
Building the operating system for autonomous cars brings not only technical challenges. Let’s imagine that the braking system fails in a car driving at high speed. Let’s say it happens close to the pedestrian crossing. The accident is inevitable. There is a fraction of a second left to save the pedestrians. What remains instinctive for human (and source of lifelong trauma) is simply the sequence of decisions for a computer, determined by the way it was originally programmed. Whose role is to decide then how it should be programmed and what decisions it should make in this kind of situations? Turn, hit the tree and kill the passengers or protect them and hit the pedestrians? Should the decision be different if pedestrians break the law and cross the road on a red light? Should the number of potential victims be taken into account? Their age? Sex? Social status? Massachusetts Institute of Technology made a social experiment and launched a website where everyone can make this kind of decisions and check how their answers look compared to others.
We may not realize that the car has become a vital part of our daily lives and we’re used to frequent social interactions with other drivers. It often happens that suburbs lack sidewalks and pedestrians and drivers share the road. Crosswalks are not designated and pedestrians feel like crossing the streets wherever they want. They usually make sure that the drivers see them and they generally expect them to slow down. They often come close to the street, wait until the fast cars pass through, and cross the road before the others arrive.
Some situations that seem obvious for us may be extremely difficult for the autonomous cars. They will avoid any dangerous situations and brake suddenly when a pedestrian gets close to the street. This, in turn, may bring a lot of traffic jams - the very same that were supposed to disappear along with the rise of autonomous cars. We will have to get used to living with new members of our communities.
The operating system of the autonomous cars and central system of maps, roads, and collision-free driving will be one of the most complicated computer systems in the world. A part of the system won’t be programmed but will be “trained” based on the great amount of data collected. As far as such programs are concerned, we simply know that they behave correctly in the situations they were tested against. It’s quite the opposite when it comes to the occurrences that never happened in which the system is totally unpredictable.
Tragical may also be the cyber attack on the central system or the single car. Hackers may want to take control of the car giving the new meaning to the “booby-trap car”.
Security instead of privacy
The rise of autonomous cars means the very end of our privacy which is already highly infringed by technology. Data from car video cameras, sensors and radars will be sent to the servers in order to enhance algorithms, update maps and store car’s behaviour for legal reasons. All the information collected could be analyzed using the existing algorithms with a face, mimics, and emotional state recognition. This, combined with information concerning our location, may be kept for further studies of behavioral patterns. As a result, we will be watched and controlled all the time. The only chance to gain some privacy will be by running away into to the countryside. Car producers and all the companies they share data with will know all about us, as soon as we leave the house or move the curtains. “Minority Report” scenario will become reality.
Many of the new technologies have an unexpected impact on our environment and social life. In the 60s, when the first network was built allowing the exchange of academic materials, no one would have imagined that it could later turn into the Internet, which has caused a crisis in journalism, changed the way we acquire knowledge about the world and influenced the results of democratic elections. The same goes for the rise of autonomous cars - it may change our lives in ways we can’t even predict right now.
Polish version of the article was originally published in Rzeczpospolita Plus Minus and it is available here