The Future of Autonomous Driving with AI

Ben Hoyt
9 min readNov 28, 2023

Technology Overview

What is the technology

The market for transportation around the world is ever changing, and each person looking for some way to make a commute easier on a daily basis. Autonomous driving has become a solution for those that might have longer drives and don’t want to pay as much attention while on the road. This technology is beginning to be expanded upon with artificial intelligence, which is the intelligence of machines and software, which can further progress the processing of this tech. The combination of the two would give a vehicle that makes decisions and drives by itself.

How it works

Vehicles that are equipped for autonomous driving have sensors, actuators, machine learning systems, and very powerful processors that assist the car with driving without human assistance. With the help of those sensors, the vehicle’s computers then knows where it is located on the road and where the cars around it are located in relation to itself.

https://www.automotiveplastics.com/mobility-trends/autonomy/

Key Features and Benefits

Having a car that will drive itself and then make decisions regarding lane changes and oncoming traffic opens up quite a few possibilities to activities while driving, as well as uses for your vehicle.

To begin with, they are helpful for people with disabilities because it allows driving to happen with little to no input. The artificial intelligence imbedded within the software of the car would be able to detect dangers and things that need to be avoided. If an elderly individual was to be put in an autonomous car and have a destination set, they would not need to worry about accidents along the way.

https://www.kenshawtoyota.ca/2016/06/14/this-is-why-toyota-wants-to-make-self-driving-cars/

Additionally, self-driving cars, when used in massed, reduce collisions and keep the road safer by removing the aspect of human error. Software is able to process and output data faster than the human mind and theoretically can react faster. It is estimated that self-driving cars could save up to $800 billion in the new lack of crash related health issues, fuel savings, more efficient transportation, and many other ways.

Key Suppliers/Producers/Developers in the Industry

The automotive industry as a whole is quite large, but the autonomous car market is a bit more niche and only has a few major players.

Tesla Inc. is a large name in the industry, providing the market with a steady flow of mass produced electric vehicles. They as well as many other brands have their own sensors and software to execute their self-driving operations. Their cars use radar and ultrasonic sensors to feed computers with data of the cars location on the road.

Another large company that is making progress in the computing area of the self-driving cars is Nvidia. They are producing computer chips for these cars that provide redundancy and fail-over safety, which further progress the level of safety the car and traverse the road with.

Technology in Context

Current and Forecasted Adoption

Being that there are multiple levels of autonomous driving, certain higher tech models are more expensive and have less of a market. Most of what is seen in todays market is considered to have “level 2” self-driving. It is estimated around 41% of new car sales are cars that have level 2 self-driving implemented within them.

For context, level 2 self driving allows for the foot to be off of the gas and brake pedal. Level 3 self-driving allows for hands-free operation. Level 4 allows you to take your eyes off the road, and level 5 is completely autonomous.

However, when it comes to the forecasted adoption of these cars, the future is starting to not believe they will work as well as intended. When doing light researching, you can see that these autonomous cars do not always make the right decisions, making drivers end up in construction zones, incorrect lanes of traffic, or even into car accidents.

Impact

With near full adoption of autonomous cars, there would be significantly less accidents on the road, as well as fewer cars since rides could be shared more easily. Since distracted and drunk driving are some of the biggest reasons accidents happen, it is estimated that self driving cars would remove 90% of the accidents that could occur.

Requirements for Successful Implementation

EDN states that to truly implement autonomous driving, all you have to do is get a car from point A to point B safely. To do this, there is a three state process.

  • The first stage is perception and how the vehicle learns and understands the environment around it, as well as understanding where it is in relation to other things on a map.
  • The second stage is motion planning, which gets its data from the perception stage and then calculates its movements.
  • The third stage and final motion is motion execution, which is where the calculations from the second stage are executed.
https://www.tomorrow.bio/post/challenges-for-mass-adoption-of-driverless-cars-2023-06-4603475547-iot

However, for successful implementation, there are many barriers that still need to be broken through for full implementation. Currently, there are limitations in software and sensor technology, limitations in vehicle-to-infrastructure communication, and many other small things that cause roadblocks.

Uncertainties

With the progression of these cars comes the need for better cyber security. There is no knowing what could happen if someone was able to hack into these cars and the damage that could be caused with that power.

There are also many uncertainties regarding the hardware and software on these vehicles. Certain sensors and software may not behave the same with differing weather conditions, and can even see interference from other devices, interfering with operation.

Assumptions

One assumption that many people have mentioned is that autonomous cars would eliminate the need for car insurance. Car insurance would still have to be a thing, being that there is still a chance there could be a car accident.

Ethics and Regulations

Risks

Although there are many benefits of these cars, there are also many risk associated with them. As stated previously, there are cybersecurity concerns, as well general technology issues that can hinder performance.

In the case of a successful implementation, we risk removing thousands of jobs in the public transportation industry. Taxis and busses would no longer be needed on the roads and that would remove the need for their drivers.

One large risk is that for many people, driving is something that they enjoy and having the car do it for them is not an option. That is something that will slow the progression of the adoption of these cars.

Most importantly, there is the risk that the self-driving cars are just not as good as we think they are. The cars could be less safe and get into more accidents than they would if we were to take control and drive the car. A statistic in 2022 stated that in the one year the cars had been on the road, 400 accidents had occurred and 273 of them were Tesla vehicles.

Ethical Dilemmas

When implementing these cars, we are taking away our ability to react to random accidents that occur around us. Currently, the software these cars have run off of algorithms that react to a predetermined and slowly learned set of possibilities. It can be argued that humans, although being the cause of most accidents on the road, are still better off reacting and making corrections to avoid accidents.

The ethics of the software themselves is also difficult because it is designed by engineers and guided by higher level employees. The dilemma here is that there is little interaction with the public in terms of communicating things that could be improved or changed to work differently.

Potential Unintended Impacts

These self driving cars, as beneficial as they may seem, could cause some damage if they are not implemented correctly. There is always a chance that AI could not predict something coming in front of the car and hurt someone, or it could be hacked and controlled by hackers and making the car do dangerous things. Another impact that it could have is increased amounts of poverty from lack of jobs in the transportation industry. Airports, taxi services, bus services, and other private services might be gone forever, along with the jobs that went along side them.

Regulations

As of right now, these self-driving cars are not extremely common in their higher level forms, and because of that there needs to be regulations to lessen the amount of potential harm these cars could cause.

The existing regulations for these cars have been slowly put in place over the last few years. Since 2017, 27 states have enacted laws that regard the on-road usage of these cars. Not many states have extended this rule to bigger trucks and commercial vehicles, but that is something that we might see in the near future. That being said, states are still slowly allowing these cars to be on the road and operate with minor human interference. The most recent state to adopt these laws was Mississippi as of 2023.

Looking into the near future of the cars, U.S. lawmakers have been working on revisiting prior regulations and legal protections for these vehicles, as well as looking at the scope and interest of the public on these vehicles. What limits progression is the producer’s ability to prove that their cars will work in all conditions and without having major errors. There are more proposals to expand upon this, but they are still in the works.

References

https://www.independent.co.uk/news/world/americas/self-driving-car-crashes-tesla-b2101763.html

--

--