A self-driving electric bus similar to the one pictured will be trialled in Western Australia this year
They’re thinking ahead in WA – A self-driving electric bus similar to the one pictured will be trialled in Western Australia this year (source: image via ABC News)

Earlier this year the New York Times published an article arguing governments aren’t adequately preparing for the inevitable arrival of driverless cars (Self-driving cars may get here before we’re ready):

Even though fully autonomous cars could be ready for the road within the next decade, only 6 percent of the country’s most populous cities have accounted for them in their long-term plans, according to a study from the National League of Cities, an advocacy and research group.

University of Pennsylvania academic, Erick Guerra, reached a similar conclusion. He found that none of the long-term transport plans developed for the 25 largest metropolitan areas in the USA take account of the potential impact of driverless cars.

Australia is no better. Melbourne’s strategic land use plan to 2050, Plan Melbourne, was published in May 2014. It quite rightly has a lot to say about transport, but has virtually nothing to say on the subject of driverless cars. An update paper, Plan Melbourne Refresh Discussion Paper, was published in October 2015 but it also neglects the issue.

There are many matters to address before self-driving cars could be a plausible substitute for conventional cars, but that’s not the imminent change. It’s much more likely that driverless trams and buses (and possibly trucks too) will be capable of operating safely and reliably on urban roads within a much shorter period.

Marginal Revolution’s Tyler Cowen is very bullish; he forecasts:

Singapore will have driverless or near driverless neighborhoods in less than five years.  But it will look more like mass transit than many aficionados are expecting.

Buses and trams are prime candidates for automation because they operate on a limited number of fixed routes. That makes it vastly easier to reliably and cost-effectively develop the complex 3-D street maps that Google’s self-driving technology uses.

Given they used a fixed guideway, trams are likely to be the first implementation. But they’re expensive to set-up; self-driving buses offer the greatest potential for coverage.

The promise of autonomous buses is considerable because they should cost a lot less to run than a conventional bus. Frequencies of five minutes or less should be plausible financially, giving Australia’s capitals the sort of turn-up-and-go “transit grid” hitherto enjoyed by much denser cities.

It would probably be harder to address some issues, like negotiations with drivers, in Australian cities than in Singapore, but there’s a lot of very practical stuff that city policy-makers should be thinking about right now.

One important issue is assessing what demands the reliance on mapping technology will make on roads used for bus routes, like this road in Melbourne’s eastern suburbs. What requirements would autonomous buses impose on the design and on-going management of variables like buildings, awnings, street furniture, signage, landscaping, utilities, traffic management hardware, and car parking?

How would it affect the way nature strips and footpaths are used? Would the operation of outdoor cafes be limited in some way? Would something small like a change (or failure) of street lighting impact on the reliability of the mapping technology? What management regime would be required given the number of different public and private players who have a role in shaping (and potentially changing unilaterally) the street?

In time buses could be equipped with mapping cameras which provide a continual update of changes in the road environment. At first though, the importance of maintaining predictability would be paramount.

And if the use of buses were to increase significantly in response to such a large improvement in service, it’s inevitable there’d be pressure to provide them with more dedicated road space and greater priority at intersections.

Policy-makers ought to be thinking about how such a grid of arterial transit routes connecting with the rail network might work and how conventional motorists might react. It would increase pressure to manage the use of road space more rationally e.g. congestion charging. On some high frequency transit routes there’d be a strong case for providing grade separation.

Policy-makers should also be thinking about the sort of autonomous bus frequencies that should apply in Australia’s relatively low density cities where driving still remains attractive for the great majority of trips.

On one hand, the density of routes and frequency/span of hours would need to be attractive enough to compete with driving. On the other hand, though, frequent buses with low load factors will be of questionable value if they increase pollutants and/or greenhouse gases.

Although it seems likely self-driving vehicles will make their first public appearance in the form of public transport, we can be very confident that in time the technology will develop to the stage where fully autonomous cars capable of navigating all streets will be a viable option.

As I’ve argued before, it’s vital that policy-makers articulate a clear vision of the role of autonomous cars (see What should we be doing now to prepare for driverless cars?). They need to take two actions.

First, they should start right now to clearly and firmly position self-driving cars as share vehicles i.e.like rental cars or taxis rather than privately owned cars.

Second, they should immediately focus on building a consensus on pay-per-use as the appropriate model for autonomous vehicles i.e. where the price of using roads is directly related to the distance and/or duration and time of use. The opportunity for a new approach is akin to the early stages of the transition from fixed line phones to mobile phones in Australia, when timed charging of local calls was introduced.

If it’s taken for granted that self-driving vehicles will simply replace conventional cars without associated regulatory changes, then they’ll very likely increase total travel and facilitate lower densities; their transformative potential to improve cities will be foregone.

There might be good news on that score. This articles suggests Google will have a commercial incentive to restrict its technology to vehicles it owns or directly controls. In the meantime, I’m going to refer to them from now on as vehicles, not cars.