Self-Driving cars/Fully Autonomous Vehicles

Aditya Rana
5 min readAug 20, 2020

Over the past few years, significant progress has been made in the field of autonomous driving technologies by several companies in the industry. Similar technologies are being used for decades in Airplanes where a significant amount of operation is automatic.

The ultimate aim for their development is to make a world where cars drive themselves, with minimal human input. This development is going to help create a world where all uber cabs drive themselves, shipping is done by self-driving trucks e.t.c.

How do these cars work?

credit: Business Insider

These cars use various cameras at different angles all around it and the computer on-board processes the images while the vehicle is in motion. The computer is taught all the traffic rules of the country and is then allowed to drive itself.

However, the whole deal is more complex than it seems. Things like figuring out which car had the right to way, managing weather conditions, and making split-second decisions is still difficult for them.

To overcome this drawback, these cars have to make use of artificial intelligence (AI) which needs to be trained with data of millions of miles being driven. The effectiveness of the system more or less depends on how much data is available for these systems.

Their possible advantages:

  • Their advent is expected to increase the overall demand for travel.
  • Less risks of accidents as they will strictly adhere to traffic rules.
  • Lower costs due to less accidents and cheaper insurance.
  • Reduced stress for commuters as they will not have to worry about parking and difficult maneuvers.
  • They will make life easier for people with disabilities.

Their possible disadvantages:

  • They may prove to be unaffordable for most people with their costs upwards of $100,000.
  • Loss of jobs as people driving trucks, cabs for Uber, Lyft will be replaced by these cars.
  • They have trouble navigating through big crowds of pedestrians, which makes them unsuitable for big crowded city roads.
  • If an accident occurs, while the car is driving itself, then who will be blamed– the software, or the driver, or the company which made the car?
  • As the cars rely heavily on their software, hacking can be a major issue.
  • In places where the roads are not marked properly, or road safety signs are non-existent

Public’s perception of self-driving cars:

A recent American Automobile Association survey shows that only 12% drivers would trust self-driving cars, while 28% people don’t know how they feel about them.

The public’s perception is more important than ever as they will influence it’s production as most of their advantages can only be realised once mass adoption occurs. As the public increasingly interacts with AVs, their attitudes toward the technology are more likely to be positive. Increasing ways the general public can experience these technologies for themselves will have a significant impact on their perception as a whole. For example, if self-driving taxi cabs become mainstream, people will be able to see for themselves how they function and the benefits of owning one.

People in the industries which may face unemployment have been particularly vocal about their negative aspects on society as a whole in order to slow down their adoption.

The skepticism:

The technology has its fair share of skeptics. Recently, the CEO of Volkswagen predicted that self-driving cars may “never happen”.

On March 18, 2018, a pedestrian was killed by a self-driving Uber. The vehicle sensed the victim 6 seconds before the fatal impact, yet it did nothing. It was programmed in a manner that it disregarded any pedestrian who intended to cross the road where there was no crossing. In addition, its near range cameras and ultrasonic sensors were not in use at the time of the accident.

Statistically, human drivers meet with at least one fatal accident for every 100 million miles driven. Google’s sister company Waymo, has driven the most out of all companies currently testing and developing the technology at around 20 million miles. Uber claims it has driven “several” million miles. At the moment, human drivers are safer than these prototype grade vehicles, although self-driving cars will become increasingly safer over time.

Current state of autonomous driving:

There are 5 levels of autonomous driving which go from 0 to 5:

  • Level 0: No Automation, cars which are driven by human input only.
  • Level 1: Cars which can apply brakes and steer, but not at the same time. These features are generally available as adaptive cruise control and lane keep assist.
  • Level 2: Cars which can do both adaptive cruise control and lane keep assist both at the same time. An example can be Tesla’s famous Autopilot feature.
  • Level 3: Cars which can only be driven autonomously under specific conditions. Although the driver has to be present behind the wheel at all times, he/she can take their hands off the wheel.
  • Level 4: At this level, cars can drive themselves for most known use cases without any human interaction. Google’s Waymo has reached this stage already.
  • Level 5: This is where cars are more than capable of being driven autonomously. There might not even be a steering wheel in such cars.

As of now, only semi-autonomous cars somewhere between level 2 and 3 are available for the general public to buy and use. These cars have some of the sensors required for autonomous driving and require human input in order to be driven.

These cars are capable of preventing collisions by braking or steering the vehicle just before the foreseen impact. In addition, they can follow the road markings to steer the vehicle while on highway and can reduce its speed while on cruise control.

Self-driving cars have a long way to go, from being safer than human drivers and ultimately replacing them.

The hurdles self-driving cars have to overcome:

Self-driving cars, at the moment, find it difficult to drive themselves at poorly marked roads. This makes them very unsuitable for areas without fully developed roads. Until carmakers can figure out a way to introduce them in such areas, people will consider them to be risky.

In addition, they will cause unemployment as cab drivers will be fully replaced by them. A way of dealing with this needs to be found. If these cars get hacked, there can be huge consequences. As these cars will be able to connect with each other, when one car gets hacked, others may be compromised as well.

Lastly, there is an urgent need for specific laws to be made for these kinds of cars. No country, as of now, has specific regulations in place for self-driving cars.

--

--