Autonomous driving – Insights from a Unicorn
Author: Carlo Nateri
Estimated reading time: 7 min
Tired of reading? Let our lovely AI Amy read it for you:
What we mean by autonomous driving
With the term “autonomous driving” we refer to autonomous vehicles, also known as self-driving cars, driverless cars, robotic cars or robocars (yes, they have quite a few denominations!). Those vehicles are capable of sensing the environment around them and therefore moving safely with little or no human input.¹
The question though is, how? In this article we will attempt to answer some questions about autonomous driving, through a deepening of the industry’s history and future outlooks and through an interview with a young expert.
History of self-driving cars
You may be quite surprised to find out that the first experiments on self-driving cars have been conducted in the 1920s, when the first radio-controlled car travelled up the bustling Fifth Avenue in NY.²
The first actual trials, on the other hand, took place in the 1950s and have continued since. Among the most important, it is worth mentioning the RCA Labs trail for General Motors: a series of detectors buried in the payment paired with a series of lights installed on the edge of the road. The car received impulses by the detector circuits which guided it and determined the presence and velocity of any metallic vehicle on the road.³
During the same years, General Motors displayed its Firebirds, a series of experimental cars that were described as having an “electronic guide system that can rush it over an automatic highway while the driver relaxes”.³
However, the first self-sufficient and truly autonomous cars appeared only in the 80s – Navlab 1 created by Carnegie Mellon University and a Mercedes-Benz designed by experts of the Bundeswehr University Munich. In such a context terminology becomes crucial: autonomous means self-governing and it represents a true revolution in the field. In fact, before those projects, cars were partly autonomous or automated, which means that they were designed and constructed to move autonomously for certain periods of time without continuous driver supervision but driver intervention was still expected or required.⁴
During the 90s the technology in the industry moved considerably forward: not only the speed of autonomous cars increased from 30 km/h to 130 km/h but also their capacity of moving longer distances grew in just a few years from 1000 km to 5000 km!⁵
Vehicles started having video cameras onboard and used stereoscopic vision algorithms to understand the environment around them.
The ParkShuttle can be considered the milestone of this period, it was the world’s first driverless vehicle. It worked thanks to artificial reference points (in easy words, magnets) in the road surface, moved by an automated people mover.⁶
During the early 2000s, the US Government funded three military efforts known as Demo I (US Army), Demo II (DARPA), and Demo III (US Army). These projects gave students and researchers the chance to investigate how autonomous cars could reduce the burden of transportation problems such as traffic congestion and traffic accidents that were becoming increasingly present in many urban areas.⁷ During these years, also the giant Google started researching and developing autonomous cars, but decided – in an initial moment – to do it privately and to put off public announcements.⁸
The first license issued in the United States for a self-driven car was to a Toyota modified with Google’s experimental driverless technology in 2012.
Additional remarkable steps forward are the 2014 Mercedes S-Class (with options for autonomous steering, lane-keeping, acceleration/braking, parking, accident avoidance, and driver fatigue detection) and the 2014 Infiniti Q50 (with cameras, radar and other technology to deliver various lane-keeping, collision avoidance and cruise control features).⁹
Furthermore, during 2014, Google announced plans to unveil 100 autonomous car prototypes built from scratch inside Google’s secret X lab. Tesla Motors, on the other hand, announced its first version of AutoPilot, followed -the next year- by a new AutoPilot technology through a software update for the cars equipped with the systems that allow autonomous driving.⁸
In August 2016 Singapore launched the first self-driving taxi service, provided by nuTonomy.
Nowadays, many popular automotive manufacturers, such as Audi, Ford, Mercedes Benz, BMW, General Motors, Toyota and Volkswagen are developing autonomous car systems.
To conclude our history of self driving cars, we will shortly summarize the 5 Levels of autonomous driving that, until today, have been recognized:¹⁰
Level 1 – Hands on – the drivers and the automated system share control of the vehicle.
Level 2 – Hands off – the automated system takes full control of the vehicle but the drivers need to be ready to intervene at any time.
Level 3 – Eyes off – the drivers can safely turn their attention away. The driver must still be ready to intervene with some limited conditions, like traffic jams.
Level 4 – Mind off – like level 3 but the driver’s attention is never required. Self-driving is possible only in certain areas or in certain circumstances.
Level 5 – Steering wheel optional – no human intervention is required at all.
So, looking at their history it really seems like in the future we will be transported and driven around by these sort of robot-vehicles! Some people are convinced that thanks to this technology driving will become faster, more productive and even safer. Some companies truly believe we will soon own one of these cars, while Uber advocates think that is unlikely, and – instead – think they will be the taxis (or better, Uber) of the future.
Or maybe not. Public opinion is truly divided on this matter, it swings between enthusiasm, faith in progress, confusion, concerns about safety and privacy. Although we’ll certainly see massive progress and many limited services and features over the next decade, scepticism remains on whether computers will be driving completely on their own or as well as humans can. Just think about how many times, while driving, we take advantage of eye contact with other drivers or with pedestrians and how many times we do unpredictable things to avoid random and potential accidents. The underlying question is: will the “cars of the future” be able to do that as well? Will they sense pedestrians and other cars’ intentions and draw conclusions from them? Of course, since we cannot be 100% sure about these fundamental aspects, we can’t help but wait for technology to run its course. Indeed, what we can do right now is embrace the help these innovations bring: such as the ability to stay vigilant far beyond our ability to do so, to react faster when unpredictabilities come at us, to avoid misjudgements (e.g when a stop sign is present as well as when one should be there but has been removed).
Interview to Federico Sanna
Federico is Italian and has been living for the last 5 years in London, UK. He has been a student, a researcher and a teaching assistant at the Imperial College, where he studied Engineering and specialized in Machine Learning, graduating with Honors.
He is currently working at Arrival, a technology unicorn start-up creating purpose-built Electric Vehicles. As the Autonomous Driving engineer, he is responsible for the correct integration, communication and operation of the sensors onboard.
Federico is truly passionate and enthusiastic about his work, thus, we decided to ask him a few questions to have a young and fresh perspective on the matter:
Federico, what is the Level technology is at now? At which of the 5 Levels of autonomous driving do you currently work on? In how much time do you think we’ll arrive at Level 5 (Steering wheel optional and no human intervention required)?
“This is a really tricky question.. Every producer is trying to raise to Level 5 autonomy, most of them pretend they have already developed the whole technology and that they are comfortably at Level 5. The truth is that defining these levels is difficult. At Level 5 the vehicles are able to drive by themselves but with what degree of safety? I would say that now we have comfortably arrived at Level 4 but very soon we will achieve Level 5 with complete safety. “
What kind of technology, or public consensus, do researchers lack nowadays to arrive at Level 5?
“Another interesting question.. This especially goes to two very different levels: the technology side and the ethical side. I would also like to add another level that goes alongside these two, the regulation one.
From a technology perspective, what is missing is reliability and the testing part. We need to demonstrate that if anyone is taking an autonomous taxi for example, you can take it a billion times without any accident happening. But, again, which is going to be the actual line? Is it an accident every billion times or an accident every million miles…?
Another thing that is lacking is cheap technology. Switching from expensive sensors such as Lidar (Light Detection and Ranging) sensors that work as scanner of the reality through lasers, to cheaper ones such as radars or – as Tesla is trying to do – cameras.
Talking now about the public consensus, at the moment most of the people are reluctant to get into an autonomous car, but here is an example that I always make related to this:
When autonomous elevators started becoming popular, people were very scared to get into an elevator that was not operated by a human technician. Once the elevator became “autonomous” let’s say, people were very skeptical in the beginning but very soon everyone transitioned to the new generation of elevators. I think that the same will happen to autonomous vehicles and that our own children will ask us how we could trust going around on a street with people driving cars and without the safety of autonomous vehicles.
To conclude with the third point I wanted to make, the governmental point of view, I think it will follow the technology improvement and it will do it quickly. The regulations will develop differently in different parts of the world. I believe that in the US the regulations will be almost immediate and then Europe will follow. “
In which field do you think autonomous driving will first spread? Military? Private? Industrial?
“The actual question here is where you think is going to be more profitable and which are the barriers to entry at the moment. If the barrier is the cost, then autonomous vehicles will spread first in the military field. If on the other side it will be affordable, then these vehicles will spread first among the private sector and the industrial sector – meant as a B2B kind of sector – . I really think that as soon as autonomous driving will break in, it will spread among any kind of field. I do not really see its application in the military field as much as I see it in the industrial or in the private sector. Not having a driver means you are not risking a human life of course, but the real aim of autonomous driving is to make transportation incredibility cheaper, much safer and to disrupt the habits of individuals for what concerns for example owning a car. For now the real barrier is the cost, and is something that is not an issue for the military sector, so for now I do not believe that autonomous driving will revolutionize the military field as much as the private or industrial one.”
Let’s wrap up
To conclude our brief insights on the world of autonomous driving, we can deduce – by looking at its history, at the next steps of the industry and at the opinion of our expert Federico – that the future of self-driving cars, with little or no human intervention, is not so distant as most people may think! So, what can we do? As single consumers, we have no choice but to start embracing the idea that in the near future our cars will drive for us (and maybe start saving some money to be able to afford this new vanguard). While, as investors, it may be smart to start thinking about investing in companies that focus their R&D on developing the Level 5 of autonomous vehicles. It could be both a way to help humankind in faster achieving new technologies and a way to put your money in a sector extremely profitable. Think about having right now the shares of the first company that in one or two years from now will design the first commercial completely self-driving car!
1.”Self-Driving Uber Car Kills Arizona Woman Crossing Street”, U.S., 2020
2.”Science: Radio Auto”, TIME.Com, 1925
3.”Wayback Machine”, Web.Archive.Org
4.”ROBOT CARS – Autonomous Vehicles – History Of Self-Driving Cars – Best Robot Car”, People.Idsia.Ch
6.”Operations Contract Of Driverless Parkshuttle Extended With 2 Years – 2Getthere”, 2Getthere
7.”Wayback Machine”, Web.Archive.Org
8.Morby, Alice, “Google Abandons Plans For Self-Driving Car”, Dezeen
9.Art, SciTech, “Autonomous Vehicles: When Will We Get There? – Scitech Patent Art”, Patent-Art.Com
10.”SAE Self-Driving Levels 0 To 5 For Automation – What They Mean – Autopilot Review”, Autopilot Review