The Case For Autonomous Shuttles

Last year, I spent a few months in the US. Anyone who has come from Europe and spent time in the US will have immediately had a bit of a culture shock regarding transportation: everything is huge! Large passenger cars, large trucks, large cupholders.

And more importantly, its urban transportation system is broken.

Now I don’t say this from an “Europe is superior” attitude — the same could be said of Europe, but for different reasons.

The large sizes of pretty much anything related to roads and transportation are the logical result of years and years of car-centric city planning. In an ideal world, the US would have developed over the years a transport structure that mixes in more modes of transport like trains, which are more environmentally friendly. However, the relatively low density and wide sprawling urban areas have resulted in what we have today: cities that need a car to be traversed, and a rail infrastructure that would need many many years to get developed to a point where it’s accessible to the majority of the population.

However, climate change and emissions goals are here and need to be addressed now — not in the 50 years that an update to the rail structure would require.

This is problematic because the solutions we have today aren’t good enough to compete with cars: bicycles can’t be used for the longer distances that most of the population needs to cover daily, and buses just aren’t flexible enough. The urban planning hellscape in some US cities means you have no way to cross the road (or rather, highway) if the bus stop happens to be opposite McDonalds.

This is where Uber came in, or at least, their initial premise. When it was announced, it wasn’t just about cheap labour and lower job security for drivers. It was meant to offer a ride-sharing service, with emphasis on the sharing.

Of course, that’s not what happened in the end; Uber turned into a VC-subsidised taxi service whose business model hinges on undercutting the costs and salaries of traditional taxis. This is largely a 1:1 service, where a vehicle drives one passenger, or a few in the best case scenario. While it is an alternative to personally owned passenger cars, it does next to nothing to reduce road congestion and pollution.

After discarding Uber, the most promising alternative are autonomous shuttles. They exist at a unique niche that can transform transportation as we understand it: they are smaller and more flexible than the bus, but when compared to passenger vehicles they can still provide an important reduction in the number of vehicles on the road.

Being smaller than buses gives them a few particular advantages. The main one is that they can scale very well to the real time passenger demand. If we have ten shuttles covering two transport routes, we can adapt on the fly all the way from a 5/5 split to a 9/1 without leaving either route unattended — particularly useful for routes to stadiums, stations, airports, or other destinations with momentary surges in passenger demand. That flexibility doesn’t exist if we serve the routes with a smaller number of large buses. Flexibility can be taken to the extreme, by allowing the passengers small detours so that they can adapt the route to their needs without straying far off a main path, similar to some school bus routes.

Traditionally, using smaller vehicles would come at a cost. Historically, businesses operating vehicles have chosen the largest platform that was practically feasible; this is how we have buses, and not microbuses, or huge haul trucks for mining operation. The reason is to keep the number of drivers at bay — not only to keep wage costs low, but because driver shortages make it hard to even find people willing to drive in circles for hours every day. Public transport exists at a delicate cost balance — any increase in costs, and passengers wouldn’t be interested in using it anymore.

By being smaller and not requiring a driver (at least, that’s the theory), autonomous shuttles can offer lower operating costs. Plus, the existing ones are already electric — being smaller means they’re easier to take out of operation one at a time to charge for a little while. A large bus needs a high utilisation rate to stay cost-effective, and the huge batteries required to have a 25-ton vehicle moving a whole day explain why we don’t have many electric buses. In contrast, a small shuttle can afford to run with only a few passengers at a time, making it comparatively easier to deploy just a few to areas where buses wouldn’t be practical.

Obviously, shuttles are not the only solution for future cities and just a way to cover some of the use cases and gaps in transportation that we are currently missing. For the longer distance, trains are great, but transforming a country such as the US to a train-first transportation network would take decades that we don’t have; they are best suited for the longer distances, and complemented by coaches that offer the flexibility that trains don’t have. For last-mile transportation, bicycles are perfect; however they are not a valid solution for people with mobility issues and they start becoming problematic after a certain distance threshold.

Shuttles fill this gap between cycling and long-range transit: they would be perfect for the trips of 3–10 miles that are typical in metro regions. Unlike bicycles, they are weatherproof, and the rider doesn’t need to get sweaty in longer distance trips!

This way, shuttles are a perfect bridge between the current infrastructure and the cities of the future, designed around people and not cars. They can leverage the existing infrastructure and reduce traffic demands, enabling us to close gradually more and more streets to private vehicles while a metro rail infrastructure gets set up. Plus, they could already offer substantial cost saving to passengers in some areas where private car ownership would stop being strictly necessary: many inhabitants of gentrified cities proclaim that cars are not necessary as public transport covers all their needs — something that doesn’t extend far from the most densely populated areas. Having this happy medium between buses and taxis would extend this area of influence and ensure than more people can live without a car.

The trickiest challenge is to deploy a technology that isn’t quite ready for prime time yet. There are a few players in this industry, but the biggest ones (e.g. Waymo) seem completely focused on robotaxis that only work as a business case if the company itself turns into a taxi operator, and even then they have a low ratio of replacement of private vehicles. For smaller players, the technology doesn’t allow fully driverless (anytime, anywhere) operation yet.

Even with technological challenges aside, self-driving companies need to build trust gradually; this can be done by deploying small pilots in university campuses, hospitals… with safety operators that can take over if needed. This allows smaller companies to deploy faster, and prove a technology that we already know can be developed, even if only a select few have achieved at a city-wide scale.

Gathering data during these trials would also allows companies to build evidence and regulators to digest and understand that technology, and create regulations that support this transition to a more flexible transit system. Instead of waiting for the train-first public transport of the future, we can start taking action now, so that we can leverage (at a low cost) the technologies that we are creating today while we build the infrastructure of the future.