How Safe Are Self Driving Cars?

Several technologies that we see today seem like the stuff of science fiction. We have hoverboards, wearable technology like Google Glass and Apple Watches, and even virtual reality and artificial intelligence are getting their time in the sun. While we are a long way from flying cars, we are seeing the very real development of the self-driving car.

From Tesla to Google, self-driving cars have been in development for years and are slowly making their way towards the consumer market. With automatic braking systems, proximity monitors, and integrated GPS hardware becoming standard, we have already taken huge steps in automotive innovation. As increasingly automated cars come to the consumer market, a serious question is being asked by companies and consumers alike: are self-driving cars safe?

The safety of self-driving cars is defined by more than just whether or not the safety system of a given car can react to emergencies. Driving involves a lot of minute adjustments to speed, the actions of other drivers, and inclement weather conditions, to name just a few. Are automated vehicles really ready to tackle all of that?

What Are the Levels of Automation?

This may seem like an oversimplified question, but there are many levels of automation, and all of them serve as steps that lead to truly self-piloting vehicles. The first two layers of automation have become standard in newer cars and are thought of as driver assistance tools, which improve safety. Power steering, cruise control, automatic braking, and lane change assistance all fall into these categories. Drivers must stay alert and focused, but they have predictive and responsive systems to help them stay comfortable and respond to unexpected changes.

The third level of automation still requires a driver, but the car is able to perform safety-critical functions. The transition between driver-controlled and automation-controlled actions is of special concern here and is the source of safety concerns in partially automated cars. If the emergency systems don’t respond, or the driver isn’t able to engage them in time, there still needs to be mechanical backups that can be used.

The fourth and fifth levels of automation are where we see a major change in the driver-vehicle relationship. At the fourth level of automation, a car can pilot itself seamlessly in most weather conditions with no input from a driver. Limitations may be put on its programming, such as not being able to pilot itself in unmapped areas, or during extreme weather conditions. The fifth level is complete automation of a car, able to function without a driver in any conditions and any location.

How Close Are We to Full Automation?

Companies from Google to Tesla have talked about releasing self-driving cars, and while the research and development are taking place, do we actually have self-driving cars yet? The short answer is “no.” Tesla’s Model S, which was released in 2020, is incredibly advanced, with self-parking, lane change assistance, and even the ability to be “summoned”, but it still relies on having a human driver to function. Tesla intends to release a fully automated car in 2021, but Waymo already has limited self-driving cars on the road in Phoenix.

Waymo is the name of what was previously the Google self-driving car project. Now, it has an app for ridesharing that specifically utilizes its automated vehicles in the Phoenix metropolitan area. Their technology has also been using machine learning to develop systems that can drive shipping trucks. While their tests have proven hopeful, they are not yet releasing a consumer car available for purchase.

The most important part of building a truly autonomous vehicle is the technology that will replace the constant calculations and adjustments that drivers make based on the sensory cues that they have learned to interpret. This is being done through the use of machine learning, a concept that is the root of basic artificial intelligence. A large amount of information is processed by advanced computer systems, from situations that have already occurred.

Using predictive algorithms, these computers can predict the outcome of new data, based on how prior data has been performed. This is similar to how humans learn from past experiences. Machine learning and artificial intelligence are seeing massive growth right now and are finding applications in medicine, security, and safety.

Physical and Cyber Safety

There are multiple angles to consider when discussing automated vehicles. The biggest concern is that right now we don’t have the structures in place for a vehicle to fully independently respond to an emergency. This becomes even more concerning considering the often delicate computer systems involved in these cars. A major concern is drivers’ over-reliance on automated systems, becoming distracted and unaware. With the possibility of greatly reduced response times, the potential damage that can be caused by an automated vehicle failing to respond can be massive.

With advanced computer systems being vital to the success of self-driving vehicles, the possibility of hacking and ransomware becomes a high-risk concern. While these can be damaging on a personal computer, when applied to cars, owners could be locked out of or even into their cars until they respond to the ransomware demands. If the vehicle is interfaced with other technology such as their phone or in-home digital assistants, hackers could have the ability to access damaging personal information as well. 

With all of these concerns around safety and potential damage, are there pros to the creation of automated cars? Absolutely. To start, the implementation of self-driving cars offers up reduced traffic jams and emissions. As vehicles respond at optimal speeds on the quickest routes, they are expected to increase fuel economy by 10% and reduce fuel emissions by 60%. People with disabilities would be able to be more mobile than ever before, and commuter times would drop significantly. While self-driving cars may not be a part of our world yet, their possibility will soon open up exciting prospects for the future of the environment and the impact that they can have on our lives.

Luke Smith is a writer and researcher turned blogger. Since finishing college he is trying his hand at being a freelance writer. He enjoys writing on a variety of topics but technology and entertainment topics are his favorite. When he isn’t writing you can find him traveling, hiking, or gaming. You can find Luke on Twitter @lukesmithwrites

Related articles

Future-Proofing Careers: Can AI Coexist with Job Security?

As you're setting out on your career journey, there's...

TV’s Evolution: Balancing Tradition and Transformation

A passionate debate regarding television's future possibilities has been...