Using driverless cars to help create liveable cities

My first blog post in this series contained a hell of a lot of subtle hints towards the topic of this. For all the knowns and unknowns about how driverless vehicles will affect travel patterns (and there are a hell of a lot of unknowns), the infrastructure.

As highlighted in this blog’s very first post, key to delivering livable cities is recognising that the needs of vulnerable highway users – particularly pedestrians and cyclists – need to take priority. Driverless cars being tested to date have some impressive claims made against their safety. A common quote is how one Google car has driven 300,000 miles without a single crash. But this is without independent verification of not just these statistics, but also things such as number of near misses with vulnerable road users. Indeed the only thing we can be sure of at present is that driverless cars will react quicker (nanoseconds) to a situation than a driver (at least 0.7 seconds), and research shows that where reaction times are slowed the likelihood and severity of a collision is likely to increase. So this technology offers the potential for reduced pedestrian and cyclist fatalities.

Whilst objectively there may be some scope for driverless vehicles to reduce the incidence of collisions with vulnerable road users, safety is also a subjective matter. The impacts of driverless cars on the subjective safety of vulnerable road users is even less well researched. What we are able to do, however, is make inferences from existing research into vehicle / vulnerable road user interactions, for which there is plenty.

Many vulnerable road user groups feel unsafe in vehicle-dominated spaces, something that is extensively backed by research. Research into shared space by the MVA Consultancy shows that high traffic flows discourages pedestrians from sharing highway space. Understanding Walking and Cycling shows the biggest barrier to take-up of cycling is safe cycle infrastructure. Making the Connections reveals traffic as a major barrier to social equality and access to services for the most vulnerable members of society. The elderly have difficulty adjusting walking pace, judging gaps in traffic, and generally interacting in more complex traffic environments. I could go on, but you get the point.

New Road in Brighton

New Road in Brighton (Image Source: Google)

To deal with this, vulnerable road users – as human beings – adopt coping mechanisms, the most notable of which is to not visit places at all, which is not condusive to good placemaking. Clearly physical seperation and priority of vulnerable road users is critical in many areas to creating quality places. But what is interesting in the case of driverless vehicles is the value vulnerable road users place on interaction with the driver, particularly in low speed environments.

It seems counter-intuituve, but current evidence shows that interaction with drivers is uncommon. Again, the MVA Consultancy shared space research is important as it records no instances of eye contact negotiation between pedestrians and drivers, with vehicle speeds, pedestrian flows, and demarkation of the carriageway influencing the propensity of drivers to give way to pedestrians. Whilst there is an inherent human value in such contact (which may boost it’s credentials as a subjective safety tool, I don’t know), it’s use as a negotiating tool appears overstated.

The current research indicates that in a driverless future, the technology may not significantly influence how vulnerable road users interact with vehicles and their subjective feelings of safety, even if actual safety is improved. This may change as the technology becomes more widespread and proven, particularly on vulnerable road user safety. But more research is needed for that.

The elephant in the room in all of this, which you caught a quick peek of earlier, is how driverless vehicles are able to support liveable cities through facilitating infrastructure change to favour people-friendly modes. Looking into the research on this, much has focussed upon how driverless vehicles interact between themselves, and interact with Intelligent Traffic Management Systems to produce optimal vehicle flows and reduce delays. Examples include space-time reservation systems at junctions, and the well-known road train experiment undertaken by Volvo. But there is no data on how much extra capacity driverless vehicles may deliver. What we can assume is that there is potential for greater highway vehicle capacity through elimination of driver mistakes, and more efficient movement of vehicles along links and through junctions. In theory.

But as we know by now, what is beneficial for vehicle capacity is not always condusive to livable cities. Lets look at the main constraint on highway capacity – junctions.

Signal controlled junction

Signal controlled junction with pedestrian crossing at Finsbury Park, London

My posts are long enough without going into the details of junction modelling. But to give a quick overview of the process

  • Count the number of vehicles on approach links and turning movements;
  • Get all of the relevant details on the junction – geometry, signal timings, topography to name a few – that affect the capacity of the junction;
  • Whack it all into a modelling package.

This then produces reams of numbers (see a Modelling Report to see how bad this can be) of which 3 are of most importance for understanding the operation of the junction: total delay minutes, mean delay minutes, and ratio of flow to capacity or RFC – simply how well the junction flows. This Ignores debates of this approach focusing on vehicles and not person movements, and the relative weighting of modes in calculations (covered excellently by Rachael Aldred), which I will cover at some point.

Needless to say that the impact of driverless cars on junctions is not well understood currently. But there is a reason for my basic introduction to junction modelling. Bearing in mind the huge variety of factors that influence junction capacity, it is unlikely that driverless cars will have a significant effect in increasing junction vehicle theoretical capacity. But driverless vehicles are likely to lead to more efficient vehicle movements through junctions. Consequently, busy junctions are more likely to operate closer to their theoretical capacity in such a way the reduces mean and total delays to vehicles.

But this is an extremely vehicle-focused way of looking at things – as is junction modelling generally. We are interested in people here, as such another way of looking at this is that if junctions do not operate at their current theoretical vehicle capacity, could the roll-out of driverless cars mean we actually reduce this theoretical vehicle capacity to its current actual use? This gives rise to all sorts of other options for creating livable cities without huge impacts on current traffic flow – longer pedestrian green phases, advanced green phases for cyclists are a couple of good examples.

I should stress here that these are just my thoughts based upon my professional intuition. Research into the impacts of driverless technology on junction operations is hardly abundant, and I expect our understanding to improve immeasurably in the near future.

The other aspect of highways are highway links. Driverless vehicles are lauded as offering potential gains in terms of highway vehicle capacity, but this assumes that the existing infrastructure remains relatively unchanged – particularly in terms of lane capacity. But as this extremely important diagram from Manual for Streets 2 shows us, that just looks at streets in terms of their movement status, and not how they should function as a place.

Again, due to the newness of the technology means there is little research on driverless vehicles and place, and this certainly needs to be researched. But it is my feeling that the principles of good street design will be the same even in a driverless future, because what we value in our streets and public spaces – spaces on the personal level that encourage social interaction (or ideal for walking and cycling in the transport planning sense) – has remained unchanged for hundreds of years. This value has even survived the rise of the motor vehicle.

On a practical note, there is the opportunity that driverless vehicles may support the reallocation of space on links by increasing capacity in the remaining running lanes, Again, the research into this is at an early stage, with varying reports of the effects of platooning (where numerous vehicles follow each other in close proximity) on highway capacities. Some initial research indicates if just vehicle sensors were used, capacity could increase by 43%, and when you add in vehicle to vehicle communications this goes up to a 273% increase.

Let’s apply these initial estimates to an example two lane single direction link near me, Bedford High Street.

Bedford High Street

As you can see, this is a two-lane single direction carriageway, 30mph with mixed frontage (including loading) with a total width of just under 7 metres.  Under the Design Manual for Road and Bridges, this gives the highway a theoretical capacity of 1110 vehicles per hour.

Applying the capacity increases cited previously, delivery driverless vehicles would increase theoretical capacity here to between 1587 and 3030 vehicles per hour. If a single traffic lane was removed for expanded footways or cycle tracks, capacity would change again from a loss of just under 200 vehicles per hour (that could easily be absorbed by the town centre network generally), to or a gain of over 500 vehicles per hour.

(I should probably state now that it has been a long-held ambition to fully pedestrianise Bedford High Street)

The implication from these back of the envelope calculations is that the roll out of driverless technology in vehicles could maintain highway capacity for vehicles, yet at the same time dramatically increase capacity for walking and cycling and give greater potential for creating people-friendly cities.

Personally I have doubts as to whether such capacity enhancements can be realised. It’s very well increasing link capacity, but if junction capacity increases are negligible then no more vehicles can be stuffed through your highway network. Additionally, movement will still need to be managed by design so as to actively contribute to the streets’ place function, attractive town centres will still need to be walkable and easy to cycle through, and highly valued public spaces will still encourage social interaction and positively contribute to city life. The only thing that will change is that machines may drive through them, not people.

Overall, driverless vehicles offer opportunities for reinventing the infrastructure of our cities in a more livable manner, and supporting livable city strategies. This can be of benefit to our most vulnerable road users, improve the overall environment of our streets. They also offer the opportunity to maintain the status quo by simply making existing operation more efficient.

My view is that the decisions that we as professionals need to make our cities more livable have not changed, and will not changed with driverless cars. They offer potential to support this approach. We just need to get away from thinking about what driverless cars can do for vehicles, and start thinking about what they can do for people.


2 thoughts on “Using driverless cars to help create liveable cities

  1. Tim Hall

    What impact do you see driverless cars having on public transport?

    I’ve seen people suggesting they’ll render buses and trains obsolete, but I’m not convinced. Suspect driverless vehicles will start to blur the distinction between public and private transport.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s