Blog by Robbert Lohmann, Chief Commercial Officer
the 2020s Self Driving Challenge: what does the upcoming decade bring when it comes to autonomous vehicles?
BLOG: the 2020s Self Driving Challenge
When we look back at the past decade, 2010 was a good start for driverless vehicles. Masdar City started passenger service on November 28th, providing passenger service without safety steward on board. Nobody was talking about ‘mixed traffic’ and ‘public roads’ yet, except maybe small teams of DARPA inspired engineers in labs and back rooms, as people were focused on throughput and added value to passengers.
It appeared to go quickly: from the first law and automated Prius (2012) to the first autonomous car (2014) and a lot more firsts since. But with the decade now complete, we are yet to see an autonomous car operating in mixed traffic… We have been told it would be here in 2 years for the last 6 years, and it will probably take another 10 years before those 2 years have actually passed. Makes you wonder whether this is what Einstein meant when he said time is relative?
When it comes to autonomous driving the 80/20 rule is in double effect: in 20% of the time, 80% of the minimum (!) required functionality is developed. Where we are almost at 80% of the functionality, now comes the other 80% of the time to develop the remaining 20% of the functionality. According to dr. Phil Koopman (who must be right given his Dutch last name) we are actually at 98% of the functionality and have been there for a while, so it appears that the last 2% might required 98% of the effort/time…
With that in mind it is the 2020s Self Driving Challenge: what does the upcoming decade bring when it comes to autonomous vehicles?
1. De(con)struction of Cities
Before starting a rant about how ridiculous cars in cities are, allow me to make one point: I love cars. I love driving. I am not a petrol head, though. I also happen to love my bikes (yes, plural: a regular one, a mountain bike and a cargo bike to go places with the kids). And I love my kids, and my kids love the bus. Most importantly, I also want to leave them living in a world and cities worth living in when they have kids and their kids have kids – and so on. Which means I believe in using the right means of transit for every trip. And that’s not the car when heading into the city…
The popular belief about autonomous cars is that smaller cars at higher frequencies are the future of our cities. Kind of like the Minority Report vision, without the vertical movement. Which means a busy intersection in the city in the future would look like this: synchronized movements at its best. Right? Right (pronounce sarcastically). Works perfectly in an animation, a little less in real life. Can you imagine what it would be like crossing a street in your city with double the amount of cars that it has today? If not, I am with you. If you can, it most likely is at a traffic light? Can you imagine the ripple effect that traffic light has to all the autonomous cars in the city? (remember, there are twice as many in comparison to today). Exactly! Dream on. Time to get real.
Cars don’t belong in cities: they destruct them. They take up parking and road space, cluttering the streetscape and making it unattractive to live, work and recreate. Making them autonomous doesn’t change this. Making them electric doesn’t change this. Making them shared, increasing the average occupancy, does! Now that’s an epiphany: our cities’ mobility is improved by having less vehicles in them. It creates more space for pedestrians, bikes, public transit, recreation, green and open spaces.
We warned about the blind faith in autonomous cars in our 2017 whitepaper, and it is now clear that we weren’t wrong. Expert city planners are speaking out and cities are quickly taking actions that are reserving specific streets for certain modes and increasing the number and size of car-low or car-free districts. They are working hard to ensure that cars don’t destruct cities, but using autonomous technologies to support city deconstruction: selective dismantling followed by repurposing of scare space. With local governments focused on pedestrian accessibility, creating bicycle friendly routes and introducing efficient public transit, the cities of tomorrow actually become both accessible and liveable despite growing urbanization. Local authorities focus on the complete requirements and how technology fits the needs of both residents, workers and travelers.
2020s Self Driving Challenge 1: build and integrate transit solutions for cities taking into account the requirements of all actors.
2. Market reset: the end of the hype
We went on record in 2018 to predict the end of the autonomous vehicle hype within the next 2 years. Duh! Big deal, you say: Gartner already had the hype cycle peaking in 2015… And you are not wrong. See the graph below for the position of autonomous vehicles throughout the years.
The conclusions that can be drawn from the above?
- Autonomous vehicles went from being <5 years away in 2014/2015 (so, they should be here today!) to being more than 10 years away in 2019 (so, not here yet for a long time).
- A distinction between level4 and level5 automation was introduced in 2018, reflecting that the hype with regards to level5 hasn’t peaked quite yet.
The latter is most likely the achievement of the marketing departments of all the autonomous car companies. And Tesla in particular as they forecasted ‘full self driving functionality’ to be ‘feature complete’ before the end of 2019 (time is relative, remember). By the way, feature complete means ‘the car being able to be autonomous, but requiring supervision and intervention at times. And it doesn’t mean like every scenario, everywhere on earth, including every corner case, it just means most of the time.’ LOL. Or Sigh? I am a marketeer, but this is marketing at another level – much lower, actually.
Anyways, marketeers of autonomous cars have managed to redefine theory by changing the hype cycle model and stretching the peak! Which is actually a heck of a job: the valuation of the company is retained, or even increased. These marketeers certainly earned their bonus. When stretching the peak from 2015 to 2018 and even 2019, the hype cycle would look very different. Kinda of like this?
It is always easy to make predictions in hindsight. In 2020 everybody has 20/20 vision about the prior decade. The complexity of making a statement on the hype of ‘autonomous vehicles’ is in the fact that it is a category name that includes very different transit modes, including cars, shuttles, trucks and delivery vans in various implementation environments. The law of the averages ensures that the placement in the hype cycle is guaranteed to be wrong.
Each mode would need its’ own placement in the hype cycle. Which would show one type of autonomous vehicle being more advanced (or less hyped) relative to the other. CES 2019 showed the future of autonomous vehicles ain’t what it used to be, with the market recognizing that ‘autonomous shuttles’ will be here much sooner than ‘autonomous cars’. Which is perhaps what Gartner was indicating with the distinction made from 2018 onwards between level4 and level5?
The hype for shuttles peaked in 2018, on July 24 to be exact, with the first IPO. The stock has plummeted into the trough of disillusionment, indicating the start of the reset of the market. With that reset ongoing, the first signs are there of the slope of enlightenment: in 2020 the 3rd generation shuttles start operating at Rivium, extending service in mixed traffic, without safety steward (!), in 2021. Which means that for autonomous shuttles the hype cycle looks like this:
Similarly for autonomous shuttles, autonomous transportation of freight (in and near ports and distribution centers) has been picking up as well. Platooning of trucks on highways is well ahead of the curve of autonomous cars. The focus for the next decade shifts away from the dream of what might ultimately be possible, to what is possible today. ‘Earth to reality: come in please!’.
What is possible today? In general automated operations in a more controlled or semi-controlled environment. We need to make incremental steps instead of trying to make an impossible jump (and break our neck, legs and everything else on landing):
- Autonomous shuttles on dedicated lanes, providing an alternative to BRT like systems;
- Autonomous shuttles at low intensity/density locations such as a business park;
- Autonomous shuttles in a controlled environment such as an airport airside (extensive study summary can be found here);
- Autonomous freight vehicles at ports and distribution centers;
- Platooning of trucks on highways;
- Advanced Driver Assistance Systems in cars on highways, a relatively controlled environment.
There are plenty of examples of what is already possible. None of them involve cars driving autonomously. Obviously this isn’t a message supported by the marketeers of the car industry…
2020s Self Driving Challenge 2: focus on what is possible now, improving transit and safety today.
3. The first commercial autonomous vehicle applications
Where the past decade was all about small, temporary demonstrations and pilots to allow us to learn, the upcoming decade needs to be about what we learned. Put our learnings into practice, by getting away from pushing technology and focusing on the requirements of the real world. We need to deploy financially viable applications. It is not about sooner or later, but about right now!
Without the ability to make the step towards commercial applications, we won’t be moving towards the slope of enlightenment. Instead we’ll fall of a cliff: cool technology, but no real applications. Google Glass anyone?
Bringing autonomous vehicles to the market is as much about technology development as it is about understanding how it is possible to make money with it. What is the added value that is being generated to people? Are they willing to pay for it? How much? And how frequently? How does it all stack up against the delivery, operations and maintenance costs? As well as the current transit service and any alternatives.
To deliver a commercially viable system today is even more difficult as the production volumes are still low and consequently the costs are still high. As a result the first commercially viable applications will not be serving low volume applications in rural areas. The focus should be on applications that feature a high volume. Preferably with a constant demand to avoid having to dimension the fleet on the peak of the demand, while not being active in the off-peak hours. To ensure the revenue from the rides fares is maximized, the transit service will need to be designed such that the trip time is minimized. Spoiler alert: that’s not in mixed traffic where the system is as fast as the slowest other road user. Dedicated lanes, whether for bikes, buses or autonomous vehicles make sense…
So, it’s time to support any initiative that leads to real applications that provide a daily passenger service. Whether they are operating truly in mixed traffic, on a dedicated lane or a segregated infrastructure – as long as the system is commercially viable. Attracting approximately 2,500 passengers per day, Rivium has been the shining example of a commercially viable application since 1999 and will feature the 3rd generation autonomous vehicle in 2020. The number of passengers is expected to increase further, especially with the introduction of the new generation having triggered a redevelopment of the business park. Now that’s how you build a business case.
Brussels Airport should start operations by 2021. Who is up next?
2020s Self Driving Challenge 3: delivery of transit systems for daily use with a financially viable business case.
The road ahead
The three challenges outlined are sure to keep us and the rest of the market occupied for the next 10 years. Let the 2020s Self Driving Challenge commence.
Disclaimer: 2getthere’s blog is a podium to share opinions and views of our industry, products and everything related to it. And being Dutch, there’s one thing we are not shy of – having an opinion. Everybody’s got them, we just express them a little more directly. OK, a lot more directly. To the point that if you aren’t used to Dutch people being Dutch, you’d think we are plain blunt. Which is not our intent. Or sometimes it actually is, but in that case we just say we’re Dutch and couldn’t help ourselves…
This Blog has been started to provide a proper podium to share our opinion, with a little bit of humor along the way. The opinion shared is that of the author, not necessarily of the company, and is obviously completely objective and should be taken very very literally. Should you beg to differ on the view expressed, please don’t hesitate to engage and share this article with your thoughts on your social media channels: if there is one thing the Dutch appreciate it is a healthy debate – no sarcasm here.