The self-sharing car

Matias Dalsgaard
Matias Dalsgaard · 31. marts 2022

In the future, no one will own a car. Cars will become autonomous robots operated by large fleet owners, and you will order a car on demand when you need car transportation.

This is pretty much the end game picture that most of us see when looking into the distant future of car technology and car usage. And when we ignore all the details and complexity of the challenges that we will face before arriving at this distant future.

As CEO of the car sharing company GoMore, I curiously follow the development of AVs (autonomous vehicles). This development will eventually have an impact on the sharing of cars, and hopefully a very positive impact. But the big questions are when and how.

Five years ago, AVs were a hot topic in DK and globally. Then the buzz leveled off for some time, because even if the topic was fascinating, reality was that cars accessible for consumers were still far from autonomous. That is still the case, but I sense a second wave of buzz around the topic now – and this time with a little more reality to it than five years ago.

That is why I some months back again decided to spend more time on understanding the current state of affairs within AVs. Where are we and where are we heading with the AVs, and what impact will it have on car sharing?

Apart from reading and studying, I this time took my private research a notch further and went to Phoenix to test the robotaxis of Google-owned Waymo. Waymo is one of the most advanced AV companies on the planet and operates a commercial autonomous taxi fleet in suburban Phoenix. Here’s a link to a number of videos from my little research trip.

It’s coming. Ish

The following is what I concluded. Knowing that future will almost always prove all predictions wrong. I’m sorry about the length of the text, but good thing is that you only have to read it if you are interested in the topic…

AVs will have an enormous impact, but not any time soon, not in general and not for car sharing. Individual AV features will be developed and improved for mass market cars. We already have many of such features: Adaptive cruise control, lane keeping assistance, parking sensors, automated parallel parking etc. Tesla is the company that has been best at marketing such features as “autonomous”, which is smart, but slightly misleading. The development will continue, and it will become increasingly safe and convenient to drive a car, but we will not see cars without drivers any time soon. The technology, the infrastructure and the legislation are far from ready.

When I talk about cars here, I mean mass market cars, i.e., the kind of cars that would primarily be used for car sharing. It will take a long time before we see cars that don’t require a driver who is able to take control over the car at least some of the time. As long as a driver is needed, the impact on car sharing will be limited. Driving might become safer and more convenient thanks to AV technology, but will still need human drivers.

What might actually be coming

I do, however, see two realistic outcomes of the progress within AVs, let’s say over the next ten years.

1) Motorway driving: The AV system might become able to take full control of the car at certain motorway distances under the right weather conditions. The motorway is a highly regulated area with all cars driving in the same direction and no pedestrians or cars crossing. The environment is fit for some level of automated driving.

2) Robotaxis: In certain restricted and well mapped and tested areas, we might see more robotaxi services like Waymo in Phoenix. In these areas, the robotaxis might be able to compete with other ride hailing or taxi services.

Both outcome 1 and 2 fall into what is called the Automated Driving Level 4. In Level 4 the vehicle is capable of performing all driving functions under certain conditions. In Level 5 the vehicle can perform all driving functions under all conditions, and the step from 4 to 5 will still be a giant step to take.

Note that I’m only talking about cars here. We will see busses and trucks travelling pre-defined routes take advantage of AV technology as well as we will see robotaxis operate in more areas in the coming years.

What is interesting considering, though, is how attractive it would really be to finally arrive at the Automated Driving Level 5 for cars. What would Level 5 realistically gain different use cases? I’m not sure that even if we could produce Level 5 AVs, this would be as advantageous as it sounds for the average car user.

The cost of lidars, radars, cameras and all the other hardware and software going into AVs is very high. Equipping a Waymo car with the AV technology (a car that is Level 4, not Level 5), costs 300,000 to 400,000 USD. This cost will of course go down, if the equipment becomes standard and mass produced. But for now, there is only a potential business case in equipping cars with AV technology if you can save the salary of a driver, or if you can sell the AVs at an extremely high price. Hence, robotaxis seem to be the most likely use case for fully equipped AVs.

Trying to think it through

But let’s assume that the cost comes down, and mass consumers can buy cars with the finest AV technology. What then? Would you want the car to do everything for you, or would you still want to take over the wheel in certain situations? I think the latter. Think of the thousands of micro-situations that involve decisions and precision that would be bothersome to articulate in a formal command. When you park, when drive up to the front door to load your car, when you enter McDrive, when you go to the car wash etc. Stop, pull over, wait here, please go a little back and park under the tree in the shadow. All such orders are easy to translate into meaningful action for humans but having to articulate all the details in a manner that would turn the orders into actionable commands for a computer-car might feel a bit tedious. Then rather just take over the wheel yourself. Also note that the Waymo cars in Phoenix do not park properly, they just stop at the side of the road to pick up and drop off passengers. They avoid the complexity of finding and choosing a parking lot.

Maybe think of it this way: Even if you could afford it, would you like to travel by taxi and give orders to a driver your entire life? I don’t think so, not all the time, and especially not if this driver is a robot that is harder to communicate with than the average human taxi driver. Giving orders can feel like more work than just driving yourself.

I could see a future, where short distance driving is not autonomous for these reasons, while you might have the opportunity of activating the car’s AV system when going longer distances. This could also mean that the AV equipment will be different for different kinds of cars. Robotaxis will get the full and most expensive AV package, whereas the mass consumer car only is equipped for autonomous driving on the motorways or highways.

But then again, on the very long run, if driving yourself simply starts feeling wrong, because this is not your primary way of interacting with a car, then you might also prefer not to drive yourself on the very short distances. The infrastructure might also be adapted to better accommodate AVs in the kind of micro-situations mentioned above. Plus, a scenario where it even becomes illegal to drive yourself – because humans are less stable and predictable than robots – could also long term be the case, especially on the motorway, being the kind of roads that are best suited for autonomous driving. But that would all belong to a very distant future.

The impact on car sharing

So, for car sharing, what does it all mean here and now? Not really anything. Cars will become more convenient to drive and safer. We will of course also transition to electric vehicles within car sharing and car usage in general, but fully autonomous cars aren’t just around the corner.

Before reaching full autonomous driving, car sharing companies might also become able to offer some automated car delivery services. Either because the rental cars will be able to drive autonomously within certain areas, i.e., be able to deliver themselves to the user within such areas, or because cars become so well equipped with cameras and sensors that they can be operated from the distance, meaning that you could have a service team delivering cars through remote driving.

That being said, a million things will happen within car sharing when it comes to improving availability, accessibility and convenience of sharing cars. New technologies and offers will be developed at large scales. That will happen regardless of the development within AVs, and this is what we work on every day in a company like GoMore. But this is a different story.

In terms of AVs, we will for a long time be talking about incremental improvements to the world as we know it. Not about radical change overnight.

If you want to read more, I recommend reading the articles from CBS News and VTPI.