Tuesday, November 4, 2014

Will Google's self-driving cars turn your city into Charleston?

Perhaps! But your vision for the city will determine that. While the patterns of sprawl can not be transformed easily, I think self-driving cars have incredible potential to change what we will build in the future.

Reducing parking spaces per capita is one of the most promising ways, but by no means is it the only way. Planners need to be aware that the potential of self-driving technology goes much further than that. If Google's venture with self-driving technology proves implementable, how we practice urban design will change in dramatic ways. The consequences on urban form are not insignificant, and the ways we move about the city could receive vast gains in energy and cost efficiencies. How we plan and code for cities will certainly change, and so it behooves us city planning folk to begin thinking about this now.

If self-driving cars become ubiquitous, this Whole Foods on Magazine Street in New Orleans demonstrates how I believe all big box stores will relate to the street in the future. This store provides no off-street parking.   (corrected per anonymous comment).

Until very recently, I was openly skeptical about the potential for autonomous cars ("robocars" for short) to do much good for urbanism. In a previous post, for instance, I reflected on the potential of robocar traffic synchronization technology to promote freeway construction. Yes, I strongly suspect that robocar platooning might encourage expansions of freeways and engineering of roads to maximize free-flow conditions for high speeds, thus keeping freeways slicing through city centers. At the very least, they will encourage the proliferation of high speed interchanges. But will North American cities therefore stay fragmented and sprawly?

In that post, I was quick to suspect that they would for a number of reasons. This was because I was highly pessimistic at the time about the potential of automated driving to significantly alter the sprawl-loving lifestyles of most Americans. I also questioned whether autonomous taxis (ATs or "aTaxis") could do this even if they have a real potential to change American consumer habits for personal vehicle ownership.

I still have some doubts about ATs reaching high market share levels, but the growing clout of Lyft and Uber, coupled with urban lifestyle changes, are developments that make me pause to reconsider. I realize that the business case for robocar carsharing (be it ATs or peer-to-peer) is currently being forged and, indeed, is proving disruptively significant. Except for sport, some are already theorizing that it could become the only way we use automobiles. Carsharing and ridesharing, as performed by companies like Uber (including their UberPool service) and Car2go, are currently paving the way to carsharing with robocar fleets. They are already building that market, only with human drivers.

So from the moment Google's self-driving cars start entering the marketplace, I suspect Robin Chase is right to believe that much of the driving population will forgo vehicle ownership in short order. That day could arrive with a speed that may surprise us. The main reasons I suspect this are the advantages robocars can provide in sheer convenience coupled with the many raw benefits of the sharing economy. Moreover, any kind of carsharing that is able to reach efficiencies of scale with wide adoption could literally blow out of the water any rival form of automobile use, and robocars, for good reasons, stand a chance of capturing that kind of level of adoption - even if it is only a partial level of adoption by most individuals using them.

For one, automated driving helps us exploit much better the lost resource represented in keeping cars empty and parked all the time. It means the vehicle doesn't always have to be parked (at least not at the place where you actually disembark from it). No coveted spots to circle around. No repetitive circuits downtown trolling for an open parking spot. A robocar has an inbuilt valet service, and while it could continue circulating without you to go park itself, that situation lends itself immediately to the cost and energy benefits of sharing. The vehicle can now go on to serve another individual if not pick up someone else at the very spot you dismounted. Parked vehicles without occupants represent wasted space and rusting metal, not to mention human time and expense. Driverless technology puts the pressure on us to capture this latent resource, and I suspect Google and entrepreneurs will quickly move to exploit it. Many have realized that this simple move will drive down the number of vehicles we will need per capita (and if coupled with empowered transit services, I suspect significantly so).

But the convenience granted in the user's experience of robocars is the overlooked game-changer.  That is the salient factor that I completely missed before!

First of all, everybody can enjoy the equivalent of "Doris Day Parking" with robocars. Like Doris Day dismounting from her vehicle at the curb in front of her covered apartment entry, you will nearly always hop off from your robocar directly at the curbside before your destination. That kind of convenience is why even really well-off people heavily use taxis in places like Manhattan in the first place: the relative inconvenience involved in storing and retrieving personal automobiles most anywhere you want to go in Manhattan is simply too much to bear.

Since the advantages of not actually having to drive a car or park it is something taxis or Uber can provide us, the extra benefit granted to you by a robocar is the fact that you never have to worry again about maintaining your driving eligibility or insurability. Think about that! If you have a driver's license, you probably take this benefit for granted, but I will call this a great advantage, since it is actually not insignificant.

But when a robocar is shared, instead of personally owned, the convenience advantages continue to pile on. As a competitive advantage to vehicle ownership, carsharing reaching the scale of ubiquitous adoption is extremely compelling and disruptive, since, think about it, you as the user no longer have to worry about owning the car, maintaining it, nor housing it. Nor do you need to stay near its parked location. Your mobility becomes completely unlinked from the automobile. What's more, you no longer have to put up with the long-term necessities of ownership, such as worrying about accommodating the near constant mismatch between the vehicle you buy and your full array of vehicle needs.

So we have the following clear advantages with carshared googlecars:

(1) The advantage of foregoing driving eligibility
(2) The advantage of foregoing vehicle ownership
(3) The advantage of tailored and atomized automobility
(4) The advantage of freeing your rents and real estate from providing automobile storage

By "tailored and atomized automobility" I mean many things which we don't typically account for as  as car owners. These include being freed from personal investment in the long-term maintenance of a vehicle. A big one we don't think about is being freed from our invested choice in one or two vehicles we can own at a time. Being locked into one or two vehicles to serve all your typical trip needs is a large burden that locks us out of a full array of automobiles to suit a particular trip need very specifically - that could include adding a utility, making some trips more luxurious, or making others more efficient and cheaper.

All four of those benefits are actually one advantage: the advantage of having your mobility completely delinked from automobile ownership.

If you own a vehicle, take a pause here to think deliberately about your life without that advantage for a few minutes. Think about all the obligations in your life to address each of its burdens, and all the particular steps they involve, not only the scheduling and the payments, but their indirect repercussions on your life choices. For example, if you commute to your office job in an SUV or pickup, think of the outright waste and inefficiency that represents. What would happen should you suddenly be liberated from each of those deficits and demands, and you discover that, hey, you have just about the same mobility with a carshared service as you would with a personally owned car? In fact, your mobility may go up, if for no other reason than you can afford your mobility better and scale it to your actual needs. Carsharing with robocars may be able to afford you this kind of liberty even in the outlying scrublands of suburbia!

Are you sensing how radically your life could be reconfigured?

Now... Lets just begin to think about the land use and urban design changes that may be in our horizon...


If we share them collectively or use them as ATs, the potential of robocars to transform our sprawl pattern is quite significant because they would dramatically lessen the need for parking spaces. That has radical implications I don't have to explain.

But there is another important consequence we should anticipate and that is the fact that the needs of retailers to capture customers will probably change greatly - in fact, I speculate this need could catalyze the most dramatic consequence of automated driving on urban form and real estate markets. What matters here is not just that the parking can go away (or at least the provision of parking near most destinations), it is how uses are suddenly reoriented to serve their customers arriving via carshared robocars. What happens when you discover that the greater portion of your customers or users is now arriving via ancy robocars, which can park themselves or be traded off between entering and departing customers as if they were a public commodity?

I think businesses are suddenly going to sense a great need to immediately front the parcel with their entries in order to receive their customer competitively at the curbside drop-off point. At last, the new urbanist street section has a compelling advantage over the strip center in terms of the one factor that really matters in sprawl: the convenience to the customer!

This Walgreens on Magazine Street in New Orleans needs no exterior signage for the "Walgreens" brand. Instead, its cosmetics section is prominently situated at the storefront. Both this Walgreens and the Whole Foods in Uptown New Orleans (above top) have realized that linear feet of frontage near the curb is the resource that is vital. When you don't need to provide parking and signs to attract customers, as in the days before the car, all that matters is what you offer as an attractive experience. No more decorated shed nor duck. There is just "a shed with delights".

In terms of the way we value property with robocar carsharing in denser areas, particularly retail and commercial property, what this means is that we will probably return to the prewar era of primarily valuing property in terms of a lot's curbside frontage. Believe it or not, the shorthand way our predecessors evaluated relative commercial property values formerly was in terms linear feet of frontage (not building price per square foot). Indeed, that's the reason we repeatedly built urbanism before the world wars. It was simply the most important factor impacting commercial property value. It was the comparatively high value we placed on street frontage that compelled people to build right to the property line without setbacks, because that was where you met all your customers and where you competed with your neighbors for them. Building to the property line maximized building value.

As carsharing grows to allow stores and restaurants to cut down and even eliminate the need to provide off-street parking, expect linear feet of frontage to commensurably become more expensive in real estate terms. Exact dimensions of storefront length, actually, will more than likely be tied to customer turnover rates at peak shopping hours. A sufficient expanse of window space to catch the passersby's attention will be valued.  I suspect stores will start to become tall and multistory as a rule, like the urban department stores of the past. The Fifth Avenue effect. Many of these anticipated effects, in terms of real estate economy, strikingly resemble the forms of the pre-automobile era of urban development!


What the carsharing that is propelled by autonomous cars will enable us to capture at significant scales is the lost usable service potential of automobiles. Presently, only taxi cab and Uber/Lyft fleets currently capture this efficiency. Carsharing increases the number of trips an individual car can serve over its usable life. With carsharing, you are, in effect, capturing more trips per net pound of manufactured goods. In gross, the efficiencies gained from carshared vehicles really will add a net total benefit to the environment, enabling the same mobility to consumers for less impact, representing less wasted energy, less material life-costs (embodied energy) and less raw material intake for the same number of vehicle miles traveled. A lot of these gains are simply plugging in to the latent capacity that our present ownership-based transport system simply locks us out of. These are thus gains that can help us offset rising energy costs if we address latent demand for cheaper mobility effectively, using, of course, transit...


Does transit go away with robocars, by the way? Not at all! I believe transit will in fact become stronger if for the simple reason that divorced from vehicle-ownership, the economic advantage of using transit as part of your daily trip routines pencils out financially.

Where one can, one will save money sharing trips with strangers. That math could be easily compared with AT/carsharing apps that will more than likely be tailored to showing you your best route and trip options (Google-style) in terms of the bottom line: the actual dollar cost of a trip. Moreover, transit will be vitally important to reducing congestion in the peak times. To prevent hordes of robocars suddenly causing gridlock in the streets (since they won't necessarily be stored near their users any longer, remember), municipalities will probably build up their transit lines to move more people in and out of the downtowns and office centers. AT fleets will correspondingly charge higher rates to prevent gridlock and to encourage modal shifts (gridlock hurts them too - especially if it is gridlock produced by empty vehicles). In that situation, AT users entering the transit market will realize that the longer they manage to stay on transit for their commutes the cheaper and more reliable life gets. I think transit will suddenly be valued politically more evenly in a toe-to-toe contest with its main subsidized rival, the freeway.

Because of the dynamic of transit mixing, I think carshared fleets will operate in home "sectors" that circulate people locally, expecting people to plug into to high capacity transit lines for the longer/cross-town/peak trips. AT companies may prefer this situation because their fleets become more manageable when most of their vehicles are circulating near one another and they can store and service their vehicles more readily in the down times (otherwise they could be eating the costs of retrieving their empty vehicles from other home sectors and far away places). So, in the future, I strongly suspect carsharing and transit will work hand in hand. They will be thought about together as one greater system, rather than our present tendency to think of them as mutually exclusive "options". This gets to the core of what I mean by the benefit of "atomized" mobility. Yes, wealthier folks could use their shared or personally owned robocars for all trips, regardless of time or distance, but even they will benefit, because they will no longer be stuck in traffic with hordes of other people with a 9-to-5 job who have a latent demand for convenient transit, but are locked into needing to store and look after their own vehicles. This is important to realize. Folks that commute with singly occupied automobiles are commuting as much to take care of their car (because it must ultimately be stored where they sleep) as they are to get from point A to point B. Robocars free them of this. It cuts an invisible umbilical that many of us don't realize is suppressing our freedom. And when it does, transit will reap the benefits.

Urban Form

Without transit in the mix to limit congestion, self-driving cars will punish uses that centralize too much in the city, meaning similar uses will have to scatter geographically. Cities like Houston and LA and North Carolina's Research Triangle, with their scattered metropolitan centers, will likely be in a good situation to ease their way quickly into wide-scale robocar adoption. Both transit and robocar carsharing are more efficient in multicenter metros. In more centralized metros, like Charlotte, robocar carsharing will actually compel municipalities to devote more resources and attention to transit network improvements lest they will mire their cores in robocar gridlock at the peak times. I anticipate robocar commuting will be possible, of course, but quite expensive in these cities since AT companies will likely use dynamic pricing structures like Uber's to mediate supply and demand. Expect to see most people commuting into and out of job centers using high capacity transit.

Since self-driving vehicles will tend to be always circulating with or sans occupants, instead of spending their time parked somewhere off the street, they will always be in the street grid swarming to the serve their clients at their destination points.  People will sense the activity of an area by noting the rates of vehicle level changes on the streets and discerning where the swarm of vehicles are gravitating to, thus sending signals to everyone about the hives of activity in the city. They will make viscerally clear the exchange of human meetings and transactions geographically. The traveling "swarms" will give us an interesting new and dynamic "psychogeography" of our cities because traffic will no longer just represent humans moving through but the thickening of human activity, corresponding to the numbers of people entering and leaving specific areas. During business hours in the middle of the day, for example, vehicles will move out of the city center to disperse into the city and then start congregating downtown again at afternoon peak time. What will be the new behaviors, land use distributions and urban pathologies that will emerge? What words will need to be invented to describe these? Urban designers should try to anticipate what they might be, and what all of this entails, sooner rather than later!

Mapping for Traffic Control

One problem limiting Google's ability to introduce its autonomous vehicles is the mapping infrastructure that is needed. But I think Google can largely crowdsource the mapping from the early adopters who will have every incentive to do so. After early adopters, cyclists will chip in, eager to create de facto "bikeways" by cartographically indicating street zones where cars better bugger off, thank you very much.  We will see an amazingly innovative period in street design using signalling and feedback from Google Maps and the manual input of robocar users. This feedback will eventually allow every street to "teach itself" how the traffic should best navigate and flow through it. Urban Design is going to get much more organic and decentralized. Google's robocar mapping could empower local constituencies, or, on the other hand, it could empower the mandates of dictatorial DOTs. It depends on who first uses the tool effectively early on. As urban designers, we need to move in quickly to ensure that we implement inventively to empower locals in the Wild West period of robocar introduction and to make sure that we demonstrate the raw of potential of allowing self-adapting, organic paradigms of traffic control to emerge. 

There's still lot's more to say on all that and more, but it will have to wait till another good evening...

To wrap up, while the new tool of autonomous driving will have many upsides, it can serve sprawl and conventional ways of doing things every bit as much as its latent ability to do great things for walkable, more congenial and humane fabrics like the Charleston peninsula. The potential to sprawlify or to make Charleston with this tool is equally there. Planners and urban designers can't be lazy. We must be ready to spring to action when robocar fleets arrive to do what can make us prosperous in a new day. So, let's start thinking harder about these potentials!