Once little more than a Hollywood-like dream, autonomous cars became a reality in the 2010s, thanks largely to research by technology companies like Google. Although they have not yet merged into the mainstream and you cannot buy one at the time of writing, self-driving prototypes travel millions of kilometers annually on our nation’s roads.
It’s not easy to compress the history of self-driving cars to less than 10 milestones, but we did our best.
The driverless dream begins
Other than that, Houdina’s name sounded similar to that of the famous escape artist and illusionist Harry Houdini that many people thought this was Houdini’s newest trick. Houdini visited the Houdina Company and got into an argument in which he broke an electric chandelier.
John McCarthy’s robotic chauffeur
In 1969, John McCarthy, who is celebrated as one of the founding fathers of artificial intelligence, described something similar to the modern autonomous vehicle in an essay entitled “Computerized Cars”. McCarthy referred to an “automatic chauffeur” who is able to navigate a public road through a “television camera input using the same visual input available to the human driver”.
He wrote that users should be able to enter a destination using a keyboard that would then prompt the car to drive them there immediately. Additional commands allow users to change their destination, stop in a toilet or restaurant, slow down or speed up in an emergency. No such vehicle was built, but McCarthy’s essay set out the mission that other researchers were to work towards.
No hands in all of America
In the early 1990s, Carnegie Mellon researcher Dean Pomerleau wrote a Ph.D. This work describes how neural networks could enable a self-driving vehicle to take raw images of the road and output steering controls in real time. Pomerleau was not the only researcher working on self-driving cars, but his use of neural networks proved to be far more efficient than alternative attempts to manually divide images into “street” and “non-street” categories.
In 1995 Pomerleau and his colleague Todd Jochem took their self-driving Navlab car system onto the road. Their naked autonomous minivan (they had to control speed and brakes) traveled 2,797 miles from coast to coast from Pittsburgh to San Diego on a trip the couple called “No Hands Across America”.
The big challenge is too challenging
DARPA announced its first major challenge in 2002. It offered scientists from top research institutions a price of $ 1 million if they could build an autonomous vehicle capable of navigating a 142-mile course through the Mojave Desert.
The challenge took place in 2004. None of the 15 participants were able to complete the course. The “winner” traveled less than 8 miles in several hours before catching fire. It was a heavy blow to the goal of building real self-driving cars. After the news of the fire hit the headlines, many believed that full autonomy was impossible.
Parking is getting smarter
While autonomous vehicles still looked like science fiction in the 2000s, self-parking systems appeared either as standard or as optional equipment (and mostly for luxury cars). They showed that sensors are almost able to cope with relatively challenging real conditions such as parallel parking in a confined space.
Toyota’s Hybrid-Prius offered automatic parallel parking assistance in 2003, while Lexus soon added a similar system to the LS, its flagship sedan. Ford launched Active Park Assist in 2009, followed a year later by BMW.
Google is looking for an answer
Google secretly launched its self-driving car project in 2009. Known as Waymo in 2020, it was initially led by Sebastian Thrun, the former director of the Stanford Artificial Intelligence Laboratory and co-inventor of Google Street View. It took a few years for the company to announce that its prototypes had driven a total of 300,000 miles under computer control without a single accident, an impressive achievement that caught many unprepared.
In 2014, an autonomous prototype without a steering wheel, accelerator or brake pedal was presented. it was 100% autonomous. Waymo is widely hailed as the undisputed leader in self-driving technology.
The big automakers are diving
Until 2013, automotive companies such as General Motors, Ford, Mercedes-Benz and BMW worked on their own self-driving technologies. Real autonomy was more difficult to achieve from a technical and legal point of view than many assumed, and the technology available to consumers in the 2010s was at best partially automated.
The first autonomous car death
Claims that self-driving technology is safer than human drivers were questioned when one of Uber’s Volvo XC90-based prototypes hit and killed 49-year-old Elaine Herzberg while crossing an Arizona road. Even though she jaywalte, the crash was considered to be completely avoidable. Police officers concluded that Uber’s security driver (who should take control in an emergency) was watching Hulu until half a second before the impact.
AI comes to self-driving cars
At CES 2018, Nvidia introduced a new self-driving autochip called Xavier, which contains artificial intelligence. The company then announced that it had partnered with Volkswagen to develop AI for future self-driving cars. The collaboration between Volkswagen and Nvidia is not the first attempt to provide AI to autonomous cars (Toyota has already studied the concept with MIT and Stanford), but it is the first to combine AI with production-capable hardware. It opens up the possibility for self-driving cars to perform better, as well as for new comfort functions such as digital assistants.
In 2020, most automakers chilled their ambitions for self-driving cars and took a far more realistic approach to developing the technology. There is not a single autonomous car for the general public, but some companies believe that they are about to crack the code. The billions of dollars and an enviable monopoly are at stake.
Tesla sells an option pack called Full Self-Driving, but its cars are in no way autonomous. The Federal Government specifically asked them to stop using this term. Volkswagen develops the technology in-house and is committed to using autonomous retro vans during the 2022 World Championship. And the German supplier Bosch has teamed up with the Mercedes-Benz parent company Daimler to place autonomous prototypes of the S-Class on the streets of San Jose. Further partnerships, projects, successes and failures will undoubtedly arise in the early 2020s.
Autonomy levels explained
While the terms “self-driving” and “autonomous” are used frequently, not all vehicles have the same capabilities. The SAE autonomy scale is used to determine different levels of autonomous ability. Here is a breakdown.
Level 0: No automation. The driver controls steering and speed (both acceleration and deceleration) at any time without any assistance. This includes systems that only send warnings to the driver without taking any action.
Level 1: Limited driver support. This includes systems that can control steering and acceleration / deceleration in certain circumstances, but not both at the same time.
Level 2: Driver assistance systems that control both steering and acceleration / braking. These systems shift some of the workload away from the human driver, but require that that person be alert at all times.
Level 3: Vehicles that can drive themselves in certain situations, e.g. B. in heavy traffic on divided motorways. Human intervention is not required when autonomous mode is activated, but the driver must be ready to take over if the vehicle encounters a situation that exceeds its limits.
Level 4: Vehicles that can drive most of the time, but may need a human driver to take control in certain situations.
Level 5: Completely autonomous. Level 5 vehicles can drive themselves at any time under all circumstances. You do not need manual controls.