In the science fiction movie, Back to the Future, which first hit the cinema screens in 1985, followed by Part II and Part III sequels in 1989 and 1990, an eccentric scientist called Dr Emmett Brown and his teenage friend Marty McFly time-travel on a flying DeLorean sports car to 2015. Futurologists who helped the movie’s scriptwriters imagined a flying car that was not autonomous, so the DeLorean had a steering wheel. When the first stories about the imminent arrival of driverless cars started appearing in the mainstream media in 2015, the timeline for self-driving vehicles was thought to be seven to 10 years — somewhere between 2022 and 2025 — followed by flying cars, without steering wheels, arriving before the turn of the decade in 2030.
Over the last seven years, readers of this column have read my take on driverless or autonomous vehicles, or robots on wheels, in a range of stories from the Internet of Things to electric vehicles (EVs). Sometimes I have been guilty of being a tad too optimistic about the timeline of such a seismic change and on other occasions I have carefully cited lots of caveats like regulatory factors which can slow down the adoption of self-driving vehicles. The reality is that the driverless riding experience is a work in progress, and completely self-driving cars without steering are still years away.
On Oct 26, chipmaker Intel Inc listed its auto chips subsidiary, Mobileye Global. The firm that Intel bought in 2017 is a key player in the advanced driver assistance systems (ADAS) as well as autonomous vehicle (AV) applications. Among its customers are Germany’s BMW and Volkswagen, Japan’s Nissan Motors, Korea’s Kia Motors and Hyundai Motors as well US auto giants General Motors and Ford Motor Co. In the four weeks since its IPO,
Mobileye’s stock is up nearly 40%, at a time when semiconductor stocks have been hammered and many investors remain sceptical about driverless cars. Mobileye offers advanced driving or Level 2 features like point-to-point driving that can be hands-off while still requiring a driver’s supervision. Apart from designing chips for advanced driving systems, Mobileye also develops its own proprietary software solutions for ADAS and AV applications.
Six levels of self-driving
I wrote about the six levels of automated driving systems in this column a few years ago. The US Society of Automotive Engineers came up with the list nearly a decade ago. Here’s a quick recap:
See also: Microsoft warns other firms of Russian-sponsored group in email hacking
Level 0. That’s like the old cars you see on the streets in your city. There is no driving automation whatsoever. The car is driven and controlled at all times by a human even if it has enhanced active safety systems installed. Fairly sophisticated features such as autonomous emergency braking, blind spot warning and lane departure warning, installed in older model BMWs and Audi, are all Level 0 automation.
Level 1 is very basic Driver Assistance. Vehicles feature a single automated system. Many luxury model cars have had basic functions like lane centering or adaptive cruise control for many years now.
Level 2 or Partial Driving Automation includes vehicles that can perform steering and acceleration. Drivers still monitor all tasks and can take control of the car any time.
See also: Microsoft, Amazon and Google are kingmakers for AI start-ups
There is also the slightly advanced Level 2+. It’s a bridge between L2 and L3 that builds upon Level 2 ADAS-like features that offer more functionality and safety potential while keeping the driver in the loop. L2+ cars have features like autopilot on the highway or urban point-to-point. Tesla’s mislabelled full-service driving (FSD) and Mobileye’s SuperVision are some examples. A good characterisation of Level 2+ is “hands free, eyes on”. Some of the high-end Ford models like Ford’s Mustang Mach-E have Level 2 and Level 2+ features.
Level 3 or Conditional Driving Automation is currently the next new thing. Level 3 vehicles can perform most driving tasks but human override is still required. Mercedes Benz’s topof-the-range S-Class and EQS models in Germany have the Drive Pilot system with Level 3 features. Once you are on an autobahn outside Frankfurt, all you have to do is click a button and the autopilot called Drive Pilot turns on. It takes charge of the steering, acceleration and deceleration. As a driver, you are then free to answer email, watch a YouTube video or make a few calls. If the car senses something amiss, it alerts you with lights and sounds so you can then take control. If you don’t do so quickly, Drive Pilot slows the car down and stops on the side of the road with lights flashing. Tesla has its own equivalent of Level 3 with its FSD.
But it is still early days for this level. It will be another at least 18 months before Level 3 features improve and get wider adoption in a range of cars rather than just a couple of Mercedes and Tesla models.
The biggest promise in recent years has been Level 4 or High Driving Automation. Vehicles with Level 4 perform all the driving tasks under specific circumstances but there is a human driver in the car which has a steering wheel, and the driver can assume control anytime he wants.
In 2018, Tesla CEO Elon Musk made an audacious announcement saying Level 4 features will be available for Tesla models by the end of 2020 which would help turn a million Teslas into robotaxis. Well, we are at the end of 2022 and Level 4 is still not a reality. If you ask Musk, he will say the Level 4 software is ready and we should have those robotaxis soon picking us up from one place and taking us to the other side of town. Musk said earlier this year that Tesla was developing a robotaxi without pedals or a steering wheel. Make no mistake, the technology is here. Some believe robotaxis will benefit most in a “controlled environment”. They still need to be fine-tuned and tested before regulators and passengers get comfortable with it. Two to three years before they get going and perhaps five years before getting enough traction for more widespread adoption.
The holy grail
Finally, there is Level 5 or the holy grail of autonomous vehicles — Full Driving Automation. Driverless cars have no steering, no driver and can operate on the road anywhere in all conditions, autonomously. No human interaction or attention is required for the vehicle to go from one point to another. Some analysts believe that Level 5 driving will first be tried in controlled environments.
Sink your teeth into in-depth insights from our contributors, and dive into financial and economic trends
Apart from Tesla, SpaceX and Twitter, Musk also owns the tunnel construction firm, The Boring Co. One use for inter-city or intra-city tunnels might be autonomous robotaxis that drive themselves around, picking you up from a suburb and taking you to another side of town. But don’t look for Level 5 cars on any street near you anytime soon.
In the aftermath of the 2008 global financial crisis which forced the US Federal Reserve to cut interest rates to nearly zero, pioneering electric carmaker Tesla emerged and attracted investor’s attention. Armed with cheap capital, Musk upended the global auto industry. Now, every traditional car maker from General Motors to Nissan and Hyundai is racing to make a range of electric car models. But electrification is just a piece of a greater puzzle in what is dubbed as transportation-as-a-service, or TaaS. The other pieces are car sharing and autonomous driving. The goal is to make an electric robotaxi or a car that you might be able to hail on your phone app.
Indeed, the entire inter-city and intra-city transportation is undergoing a transformation as electric, shared and autonomous vehicles take hold. UBS estimates that the global EV market will exceed US$1 trillion ($1.4 trillion) in 2026, or about 10 times the size it was just two years ago. The ride-hailing market is projected to nearly triple over that same time frame to US$545 billion. And US investment bank Morgan Stanley expects electric sky taxis, or electric vertical takeoff and landing (eVTOLs), which are still several years away, to eventually grow to become a US$ 1 trillion market by 2040. The flying cars are seen initially only as a replacement for short-range flights for the wealthy but will get broader adoption as service gets more affordable.
A number of tech companies and a few traditional automakers are working on their own autonomous vehicles. The oldest of the self-driving project is Google’s Waymo which runs a small self-driving taxi service in Phoenix, Arizona. For now, Waymo taxis have steering wheels and human supervisors who are able to immediately take control if something is amiss. Waymo began testing a similar service in San Francisco a few months ago and expects to launch a robotaxi service in Los Angeles soon. Online news websites Axios reported recently that California has given Waymo the go-ahead to transport passengers in San Francisco without anyone behind the wheel. The Google robotaxi subsidiary still needs a separate permit to begin charging fares as it does in Arizona. Waymo has also teamed up with China’s Geely Automobile Holdings to launch a dedicated robotaxi service with its subsidiary Zeekr.
General Motors owns a self-driving car firm, Cruise, which has been testing autonomous taxi service in California. It plans to expand its robotaxi service to Phoenix, Arizona; and Austin, Texas next month.
But others are exiting long-gestation projects like self-driving cars. In early October, Ford and Volkswagen pulled the plug on their autonomous vehicle joint venture, Argo AI. “Profitable, fully autonomous vehicles at scale are a long way off,” Ford CEO Jim Farley said after Ford announced the news. Another AV delivery start-up, Nuro, last month laid off 20% of its workforce. Some tech giants plan to enter the fray. iPhone maker Apple has its own Project Titan which is developing autonomous vans while Amazon.com is funding its self-driving robotaxi project Zoox.
China, the global leader in EVs, plans to keep its lead in the EV space as well. Baidu unveiled its Apollo RT6 — the sixth generation of self-driving EVs built for ride-hailing — in July. It is poised to start plying on Chinese roads late next year.
A lot has changed over the last few years. For one thing, we are no longer in the near-zero interest rates world. The terminal rate, or the peak spot where the federal funds rate comes to rest before the US Federal Reserve starts trimming it back, is now 5.25%. By the time the Fed is done raising rates in the first quarter of next year, the cost of capital would have gone from zero to 5.25% in just 12 months. That is the fastest and steepest rise in cost of capital in recorded history.
When capital is too cheap, everything can get funded — even wild dreams and imaginations. When capital is as expensive as it has been in nearly three decades, long-gestation projects struggle to get traction because few are willing to fund them. Technologies like self-driving vehicles are not being prioritised by venture capital firms, bankers and investors the way they used to be. Cash-rich giants like Google, Apple and Amazon will still be able to fund their AV projects but others will have to pare their ambitions.
Assif Shameen is a technology and business writer based in North America
