Reimagining our Roads with Self-Driving Cars
Ali Attari
November 18, 2016
The last few years have proven to be exciting when thinking about the different technological innovations that have been unveiled. Smartphones, tablets, and Netflix are only some of the creations that many of us use regularly in our daily lives. Tech experts haven’t slowed down this trend, and have promised us many more thrilling inventions such as drone delivery and mind-controlled PCs. Few things, however, are receiving as much attention as autonomous cars, which provide the same freedom and convenience as the traditional automobile, but without the need for anyone to drive it.
High levels of traffic congestion and alarming rates of automobile fatalities (particularly in the developing world) suggest that there is a significant opportunity for a transportation revolution. Many cities are shifting their focus away from building additional roads to encouraging more sustainable forms of mobility such as walking, cycling, and public transportation. At the same time, however, the technology for autonomous (also known as driverless or self-driving) cars is also being developed with tremendous speed and determination.
Automobile manufacturers Tesla and Ford have been leading the pack in producing autonomous cars, but each has a different approach to producing them. For instance, Tesla is unique in that it allows its customers to test some of the upcoming features that they want to include in their cars by releasing beta versions that allow it to gather data and feedback from its customers. Unfortunately, this perk backfired when one of its customers had a fatal accidentwith a semi-truck while beta testing its autopilot mode, which offers ‘level 2’ autonomy. This level of autonomy gives drivers partial assistance with various car features such as steering and/or acceleration and deceleration using data from the surrounding environment, while requiring the user to remain attentive in order to take over control of the vehicle if the need arises. In the case of this accident, however, neither the driver nor the sensors noticed a white semi-truck against the brightly lit sky.
The differing levels of autonomy refer to how much intervention is required from the driver, with level 0 giving the driver complete control, and level 5 not requiring any human intervention, except for setting the destination. Tesla is gradually approaching the release of a level 5 autonomy vehicle, unlike Google, whose subsidiary company ‘X’ has been engaged in developing driverless car technology since 2009, and aims at directly launching level 5 cars once the technology has reached its maturity. Google's data gathering schemes are slightly different than Tesla's as well. Tesla relies on customers who are using its autopilot mode in different cities around the world, while Google gets its data from lab simulations and a small autonomous vehicle fleet roaming a few US cities.
Even though the idea of complete autonomy is still not fully realized, vehicles have incorporated assisted driving features such as traction control and cruise control for decades. Many of these features were included as partial remedies that address safety concerns. Automakers have shown great progress when it comes to improving vehicle efficiency and safety, but that has not been enough to effectively mitigate errors in human judgment. One of the greatest advantages of driverless technologies accordingly will be to reduce the number of auto crash fatalities, which claim around 1.3 million lives worldwide each year. Such cars are equipped with a myriad of sensors that continuously scan their surroundings, and are programmed to adhere to traffic regulations (watch this video for a closer look at how they function). The margin of error for a properly functioning autonomous car would only be a small fraction of that of cars operated by humans.
In addition to increased safety, there are tremendous economic benefits to be derived from the switch to autonomous cars. They can quickly assess their surroundings and communicate with each other, therefore minimizing the need for physical interventions such as traffic lights, and avoiding traffic congestion, which often arises as the compounding result of drivers’ inattentiveness. This is clearly demonstrated in the ‘jamiton’ or ‘phantom traffic jam’ experiment, which shows that humans will always bring about unnecessary disruptions in traffic, as with sudden breaking.
Although it remains early to know with certainty the exact economic and overall urban impacts of the spread of autonomous cars, initial indications are encouraging. A 2014 analysis in the MIT Technology Review estimated that the aggregate financial benefit of converting the bulk of the U.S. fleet from conventional to autonomous vehicles would be more than $3 trillion a year—an incredible 15% of the country's current GDP. This is primarily because the time that otherwise would be spent driving could be spent on other tasks or on leisurely activities. In addition, those who currently drive would be freed from that task, and can spend their time in the car more productively or in a more relaxed setting by engaging in a variety of activities such as reading or watching a video, or simply doing nothing.
Automated vehicles also can function as a force for social equity as they would enable unprecedented mobility for marginalized groups whose access to automobiles was previously limited by either physical or financial barriers, as with the elderly, children, the physically and mentally impaired, as well as those who simply cannot afford to own a private car. Such cars after all do not need to be individually owned, but can be part of the car sharing economy, as people would only hire them on demand rather than owning them.
It would be one-sided, however, to discuss these social and economic benefits without discussing the negative impact that this technology will have on many of those engaged in the transportation services sector, such as taxi drivers. Although the spread of autonomous vehicles will create jobs relating to their maintenance, programming, and management between different users (for shared ones), many people also will lose their jobs as autonomous vehicles become more common. The resulting plight of those negatively affected will need to be addressed.
Another possible liability of the reliance on this technology is that their control systems can be hacked. These vehicles are of course connected to the internet and require a constant data stream in order to operate. Research hackers in fact have managed to hack their way into a Tesla model S and a Jeep Cherokee in an experiment that aimed at testing and improving their systems, and exposing the vulnerabilities inherent in this sort of technology. They were able to obtain a level of control that can certainly put the driver and others in serious danger if they had malicious intentions. In the case of the Cherokee, the hackers managed to take over complete control of the car; in the case of the Tesla, they managed to disable some of the sensors and establish control over some of the car's features.
There are clearly significant risks to be considered and evaluated for both the individuals using autonomous cars and the public at large. The vision for many automakers and ride-hailing companies such as Lyft and Uber depends on the successful development and widespread implementation of this technology. Lyft co-founder John Zimmer predicts that individual car ownership will come to and end by 2025 in major US cities. For ride-hailing companies, autonomous driving technologies allow for a more efficient employment of their fleet. Privately-owned cars are driven on average only 5% of the time, and are parked for the remaining 95%. Autonomous cars theoretically may be used around the clock, thus reducing the number of cars needed to meet the transportation needs of a given community, and even allowing owners to generate income without the need for a driver to be physically present in the vehicle.
Such a decrease in car ownership brings with it many benefits and possibilities. One of them is an improvement in air quality as cars will be used more efficiently. Moreover, the massive amount of space dedicated to parking (in US cities, for example, each car has is said to have an average of three to four parking spaces) can now be reclaimed as public spaces for pedestrians, businesses, and cultural activities.
A country such as Jordan may not be very hospitable to autonomous vehicles because of poorly-designed and maintained roads as well as chaotic driving behavior, not to mention the resistance of many involved in the conventional transportation sector. Data gathering also needs to be improved since interactive maps that aid in directions are often misleading in Jordan due to inaccurate or incomplete data. All these will need to be addressed in order to prepare the country for autonomous vehicles. They eventually will come. We do need to prepare for their arrival and embrace their benefits, but also mitigate their disruptive impacts.
My thanks go to Mohammad al-Asad and Ashley Harris for their feedback in developing this article, and to Yasmine Abuzeid for creating the image used in it.