20 More Years? The Complexities Delaying The Arrival Of Autonomous Vehicles

By John Oncea, Editor

While progress is being made in the commercial availability of fully autonomous self-driving cars, there are still barriers that need to be addressed before widespread adoption can happen.

Autonomous vehicles (AVs), more commonly referred to as self-driving cars, have been five years away for at least 15 years. Heck, an early version of an AV was proposed by none other than Leonardo da Vinci in 1478.

Da Vinci’s three-wheeled cart was powered by coil springs and featured programmable steering which was achieved by arranging wooden blocks between gears at pre-set locations. When the brake was released, the cart propelled forward and proceeded either in a straight line or by making turns using pre-set angles – whichever path was programmed.

“Da Vinci’s cart design was so ahead of its time that its exact workings baffled scholars until late in the 20th century,” writes da Vinci Inventions. “But, in 2006, Italy’s Institute and Museum of the History of Science in Florence built a working model based on da Vinci’s design and, to the surprise of many, the cart worked. Some experts even noted that it looked similar to the Mars Land Rover.”

Work continued on AVs over the years including Francis Houdina’s attempt to run a radio-controlled, full-size automobile through the streets of New York in 1925. According to the New York Times via Jalopnik, “A loose housing around the shaft to the steering wheel in the radio car caused the uncertain course as the procession got underway. As John Alexander of the Houdina Company, riding in the second car, applied the radio waves, the directing apparatus attached to the shaft in the other automobile failed to grasp it properly.

“As a result, the radio car careened from left to right, down Broadway, around Columbus Circle, and south on Fifth Avenue, almost running down two trucks and a milk wagon, which took to the curbs for safety. At Forty-seventh Street Houdina lunged for the steering wheel but could not prevent the car from crashing into the fender of an automobile filled with cameramen. It was at Forty-third Street that a crash into a fire engine was barely averted. The police advised Houdina to postpone his experiments but after the car had been driven up Broadway, it was once more operated by radio along Central Park drives.”

Fourteen years later, at the 1939 New York World’s Fair, “Industrial designer Norman Bel Geddes showcased a dazzling array of visionary transportation concepts,” writes Mobileye. “His ‘Futurama’ exhibit featured rooftop helipads and interstate highways – as well as semi-autonomous vehicles to travel along them, using a combination of radio control and magnets embedded in the pavement.”

Enter Academia, Then The Military

As the years rolled by, institutions of higher learning began driving AV development, Mobileye writes. First up were researchers at Stanford University who developed a small cart to navigate on the Moon using a basic form of computer vision in 1961.

Then, in 1971, “The U.K.’s Road Research Laboratory showed off a retrofitted car with auto-driving systems, the idea being to merge the features and convenience of highway driving with the safety of railway travel,” writes Car And Driver. Six years later mechanical engineers at the University of Tsukuba in Japan created the first truly autonomous vehicle, able to track white street markers at speeds up to 20 miles per hour.

“In the 1990s, universities and research institutes began to focus seriously on creating cars that could drive themselves,” Car and Driver writes. “No Hands Across America was the name of an ambitious project completed in 1995, in which Carnegie Mellon University’s Navigation Laboratory, or Navlab, successfully took a modified 1990 Pontiac Trans Sport on a 3,100-mile cross-country journey from Pittsburgh to Los Angeles, of which 98.2 percent was autonomously controlled. The car wasn’t completely controlled by the computers, though: While it used autonomous technology to steer, a human driver operated the brakes and a hand throttle.”

Around the same time, the Eureka PROMETHEUS Project was bringing together a consortium of universities, automakers, and tech companies in what was the largest R&D project ever in the field of driverless cars. “Funded by (mostly European) governments … the project culminated in 1994 with (an approximately) 620-mile drive on Parisian highways that included operating hands-free, in convoy, automatically changing lanes while tracking and passing other vehicles,” Mobileye writes.

The U.S. Department of Defense – through its Defense Advanced Research Projects Agency (DARPA) – jumped into AV research in the 1980s, starting by funding the Autonomous Land Vehicle (ALV) project which employed an early form of LiDAR to navigate off-road.

DARPA – which also supported Carnegie Mellon’s Navlab – opened the AV floodgates “with the DARPA Grand Challenge – a series of competitions that saw Stanford and Carnegie Mellon (among others) sparring for AV leadership,” writes Mobileye. “None of the contestants finished the first desert competition in 2004, but five did the following year – led by Stanford’s Volkswagen Touareg (using Intel processors). Two years later, Carnegie Mellon won the urban competition with a specially equipped Humvee.”

Next Up: Private Industry

DARPA’s competitions demonstrated the feasibility of self-driving technology, motivating many startups, technology companies, and automakers to initiate their AV development programs. This includes, according to Car and Driver:

• The successful completion of the VisLab Intercontinental Autonomous Challenge — a 9,900-mile, 100-day intercontinental test in which the vans traveled from Parma, Italy, to Shanghai, China in 2010.
• VisLab’s successful use of driverless cars to navigate the streets of Parma in 2013, avoiding pedestrians and following street laws. In many ways, this accomplishment was more impressive than the 9,900-mile trek as the cars faced far more complex driving conditions.
• Google’s entry into AV development in 2009 which started by modifying Toyota and Lexus vehicles with AV hardware. By December 2014, it rolled out the first of its own fully autonomous cars.

Other key milestones that tell the story of AVs, according to GlobalData, include:

• 2015: Tesla’s unveiling of Autopilot, Baidu driving a retrofitted BMW 3 Series autonomously in Beijing, and Delphi AV completing a coast-to-coast drive.
• 2016: Uber’s test of autonomous taxis with a human co-pilot in Pittsburgh.
• 2017: NuTonomy’s test of self-drive taxis in Singapore.
• 2018: Autonomous lorries tested on highways in the U.S. and U.K., Nissan-DNA’s self-drive experiment in Yokohama, and Waymo’s start of limited commercial robotaxi service in Chandler, AZ.
• 2020: GM launches Origin robotaxi with Lyft while Volvo and Uber announced plans for robotaxi service with Geely and Nvidia.

George Jetson Was Full Of Shit

If The Jetsons taught us anything, we’d be living in homes raised high above the ground on adjustable columns, sassy robot maids would be picking up after us, and we’d have flying cars.

Where are the flying cars?

I guess with The Jetsons being set in 2062 there’s still time for the promise of flying cars to be realized, but it sure does feel like it’s well past time for AVs to be here, doesn’t it? Predictions made nearly a decade ago promised we’d be riding in AVs by now, but proponents underestimated the complexity of developing AVs.

While significant progress has been made, the last 20% of technological refinement is proving to be increasingly difficult, writes The Conversation. For example, existing sensor technologies such as LiDAR and radar have limitations in certain weather conditions (heavy rain, snow, fog) or complex urban environments making the development of sensors with improved capabilities essential.

Edge cases – rare or unpredictable situations encountered on the road – remain a challenge as well. Autonomous vehicles need to navigate unique scenarios that might not have been encountered during training and work still needs to be done to make that possible.

Safety and reliability are concerning as well. Ensuring the safety of autonomous vehicles in all situations is a complex task and AV technology needs to be continuously improved to ensure the ability of it helping to prevent accidents and respond appropriately to unforeseen events. Also crucial is achieving high levels of system reliability and redundancy and failure to do so can have severe consequences.

Additional challenges and obstacles run the gamut from technological to psychological. These include:

• Mapping and Localization: AVs heavily rely on high-definition maps for localization and navigation. Updating and maintaining these maps in real-time to accommodate changes in the environment (construction, road closures) is a challenge. So too are GPS signals which can be inaccurate in urban canyons or areas with tall buildings. AVs are going to need alternative localization methods to operate effectively.
• Human Interaction and Trust: Developing autonomous systems that can understand and predict human behavior is challenging. Interacting with human drivers, pedestrians, and cyclists naturally and safely is crucial for widespread acceptance. Building public trust in autonomous technology is essential, so much so that convincing individuals to trust self-driving cars with their safety may be more difficult than any technological challenge. A big part of earning that trust will be the ability to program AVs to make ethical decisions in complex situations. Determining how vehicles should prioritize the safety of occupants versus other road users is a subject of ongoing debate, as is determining liability in the event of an accident involving AVs.
• Public Perception and Education: There is a need for public education to increase awareness and understanding of AV technology. Misconceptions and fear can hinder acceptance and adoption.
• Cybersecurity Risks: AVs are susceptible to cybersecurity threats. Securing the vehicle's communication systems, sensors, and control systems from hacking is critical for ensuring the safety and integrity of the technology.

Other stumbling blocks, according to McKinsey & Company, include regulatory hurdles as governments work to develop the necessary safety standards and legal frameworks, as well as semiconductor supply chain issues and the high costs associated with maintaining autonomous vehicle infrastructure. Experts now think these challenges will be met around 2035.

That said, some AVs are being tested in select cities but are still confined to geofenced areas and have technological limitations that prevent them from operating freely on public roads. Experts suggest it will take until 2025 before Level 4 AVs begin to be offered for commercial use in urban areas, The Conversation writes.

Can You Hang On 20 More Years?

Addressing these barriers requires collaborative efforts from technology developers, policymakers, regulatory bodies, and the public. Ongoing research, testing, and iterative improvements will play a crucial role in overcoming these challenges and realizing the potential of fully autonomous self-driving cars.

While self-driving technology has come a long way, the road to fully autonomous vehicles is still winding, with challenges yet to be overcome. Patience and continued research are essential as we navigate toward a safer and more efficient transportation future. We can expect to see incremental progress in AV capabilities over the next several years but a truly driverless, Level 5 AV that can operate in any environment without human intervention is still likely at least 15-20 years away from becoming a mainstream reality.