Autonomous Cars: Are We There Yet?

The monumental growth of AI has equated to huge progress in many sectors of the economy and ideas that had seemed otherworldly have now come to fruition. One of those ideas is autonomous cars. But what is an autonomous car? 

To put it simply, it is a car that is self-driving and requires very little to no human input. (Bear in mind this is different from a car that runs on an automated system, which is heavily dependent on user input and is very restricted in the tasks it can perform.) So, when it comes to building autonomous cars, the process involves a whole new level of complexity and ingenuity that many giant car manufacturers are looking to get right. 

One car manufacturer that is leading the autonomous car industry is Waymo, who have been making self-driving cars since 2014. The company drove its first fully driverless car on public roads in Austin, Texas in 2015, the car had no steering wheel, nor floor pedal and thus began its journey. Waymo’s cars do follow the conventional features of an autonomous car such as having LIDARS (Light Detecting and Ranging System), cameras and radars at every side of the car. However, as Google’s sister company, Waymo’s substantially greater financial backing allows for a huge amount of driving simulations to take place - meaning the car retains lots of training data and is able to solve a real-world problem.

A level 4 Waymo car. Image credit: Google

Yet solving these real-world problems boil down to the level of automation the car has. In the world of self-driving cars, the levels go all the way from level 0, where the car has no automation at all, to level 5, in which the car has no restrictions whatsoever and is fully autonomous. Currently, Waymo’s cars are at level 4 because even though they do not require a driver, the full automation only works in specific geographic locations and there will most likely be a speed limit for the car when it is self-driving. The area restrictions are put in place because it is advantageous for cars to drive in dry weather - optimising the chances of safety.

Being safe and not causing other drivers harm, is probably the most important aspect to consider when building a self-driving car. However, this task requires a lot more than technological power as companies have to consider ethical and moral reasoning when programming the AI for the car. This causes unavoidable questions to arise when an accident does occur. Why didn’t the car stop? Was the decision the car took morally correct? Was the accident inevitable? And since morality is subjective, each driver may not take the same decision their car would take, which could lead to major controversy between the public and car companies. This could mean companies will delay or even prevent the large-scale production of autonomous cars.

 As for the current landscape of self-driving cars, it’s looking pretty barren, many companies including Waymo have suspended vehicle testing due to the pandemic and Ford have even postponed self–driving service until 2022. Despite the pandemic, critics have said that the predictions for when the cars would come out were too optimistic and they will require at least a couple of years to be released. Tesla for example, with its renowned Autopilot feature is still far off full automation and will require many changes until it reaches that stage.

So, it does seem like we still have plenty of hurdles to cross until we reach the level of full automation, but with the ongoing simulations and real-world practice cars are having, that stage doesn’t seem far from attainable.