Explain a Technology - Self-driving Cars

While attending a family reunion, the conversation went from the latest town gossip to new cars which devolved into a heated debate about self-driving cars. One of my aunts stated, “Did you hear? A few weeks ago, another person got killed in a self-driving car. How do those death machines even work?”. Of course, being the only tech person in the family, all of their questions were directed to me. After answering their questions to the best of my ability, the general consensus was that this new technology shouldn’t be trusted and was just going to result in people getting hurt. But, is this really the case? How do self-driving cars actually work and are they safe or deadly?

Levels of Autonomy
Self-driving cars are fully autonomous vehicles. It's important to note that most vehicles are semi-autonomous with features such as self-parking. There are several levels of autonomy which includes:

• Level 0: An automated system has no control but can notify the driver of hazards.

• Level 1: Control is shared between the driver and the system. This is present in most vehicles with Advanced Driver-Assistance Systems (ADAS).

• Level 2: The system can take full control of the vehicle, but the driver must be able to take control if the system fails.

• Level 3: The system has full control and passengers can safely move their attention to other tasks besides driving, but they must have the ability to intervene.

• Level 4: The driver can safely move their attention to other tasks besides driving, however this is limited to certain “geofenced” locations and controlled environments.

• Level 5: No human actions are needed.

Sensors and Wireless Communication
Self-driving cars utilize several different technologies such as radar, ultrasound, cameras, and radio connections. These technologies are used together to pilot the vehicle. For example, Tesla’s self-driving feature called “Autopilot” uses eight different cameras for 360-degree visibility and 12 ultrasonic sensors and a front radar for analyzing the road and possible hazards. While driverless cars are not very popular at the moment due to development, but their usage is planned to increase with the development of 5G. 5G is planned to be close to 1000% faster than 4G LTE and will allow for driverless cars to quickly communicate with each other through vehicle to vehicle (V2V) communication. This also includes communicating with traffic lights and cars that are not driverless as more and more devices become a part of the Internet of Things (IoT). Connecting with objects, such as traffic lights, is known as vehicle to infrastructure (V2I) communication (Landmark Dividend LLC, 2019).



V2I is being utilized with areas that driverless cars struggle with, such as parking. Parking spaces can be given their own sensors and mapped out. This parking map can then be used for the vehicle to reserve its spot and plan for where it needs to go, that way multiple driverless vehicles won’t have conflict over the same parking spot. Self-driving cars also utilize vehicle to pedestrian (V2P) communication. One of the biggest concerns with driverless vehicles is that it needs to be able to detect pedestrians and to be aware of their exact location. It is estimated that over 15 people die every day in accidents, and while driverless cars may not be able to eliminate all accidents, it is theorized that these vehicles would be able to greatly reduce this number. So, how does a driverless car detect a pedestrian quick enough to react? For some cases, the vehicle will rely solely on the vehicle’s sensors, however, since most people carry devices that can connect to the internet, this can be used to detect their movement. Driverless vehicles use the GPS in these devices to determine a person's location near the vehicle which will give the vehicle the ability to think dynamically and, hopefully, make the roads safer (Landmark Dividend LLC, 2019).

Safety

The hatred for driverless cars began when a pedestrian was killed in 2018 in Tempe, Arizona. One argument supporting driverless cars is that most accidents are due to human error such as distractions, anger, or fatigue. The accident in Arizona was partially human error, since the vehicle was not full autonomous, most like Level 3 or 4 autonomy, and the driver was distracted. Unlike human drivers, driverless cars follow a set of rules while human drivers can be unpredictable. The main safety issue is the transition period where driverless vehicles will share the roads with human drivers. In 2016, a truck hit a Tesla car which was in its “autopilot” mode, which killed the car’s driver (Delony, 2019). Of course, there are some ethical dilemmas that have been brought up, such as "if the car can not stop in time and there are two groups of pedestrians, who should the car hit?" This issue has created several questions as to how the vehicle should handle certain situations, however, as the technology evolves this should become less of an issue.

Currently, most of the self-driving cars on the road are for testing, however there are some that have been sold to consumers. It is predicted that mass sales will begin in 2020 and it is expected that there will be over 10 million on the road by 2023. Despite these facts, people still don’t trust driverless cars with over 56% of people not wanting a driverless car and 76% wanting a brake that the passengers can engage. This stigma may remain until the usage of driverless cars becomes more widespread.

References

Landmark Dividend LLC. "Self-Driving Car Technology: How Do Self-Driving Cars Work?" Landmark Dividend LLC, 2019, www.landmarkdividend.com/self-driving-car/.

Delony, David. "Are self-driving cars safer than cars driven by humans?" Techopedia, 2019, www.techopedia.com/are-self-driving-cars-safer-than-cars-driven-by-humans/7/33437#targetText=The%20data%20seems%20to%20suggest,are%20caused%20by%20human%20error.&targetText=Self%2Ddriving%20ca.