Parking solutions are often presented as if they are some sort of universal panacea for urban traffic congestion, alongside the often-repeated claim that 30 percent of urban traffic is created by drivers looking for a place to park. Let’s leave aside the rather dodgy provenance of this data point, which I think originates in an article by parking maven Donald Shoup. There are theoretical grounds for doubt, too; after all, the ‘predict and provide’ model of traffic planning and road building is now widely questioned, since traffic actually seems to expand to fill however much road is provided. Smart parking amounts to increasing the number of parking spaces, because providing better information improves the fill factor of existing spaces. At least some research seems to indicate that the best way to reduce urban congestion is to reduce the number of parking spaces. That makes intuitive sense, too; one of the reasons why I almost never drive into central London is my well-founded expectation that I won’t find anywhere to park.

That said, I’m somewhat impressed by the latest implementation of smart parking by Telensa, a UK company providing hardware and software for Low Power Wide Area networks in a range of machine monitoring and control applications. Telensa is providing smart parking in Moscow, and has recently added St Petersburg and a Chinese project to its list of deployments. The Moscow context has been a bit of Darwinian nightmare in which the city council has burned its way through five suppliers of smart parking solutions. Telensa’s kit, deployed by its telecoms provider partner Gorizont, sits alongside sensors rolled out by an earlier “winning” consortium, World Sensing and Sigfox, who announced their own win at the beginning of 2014.

Startup office

The Telensa PARKet solution involves parking bay sensors about the size of a baked bean can, which are pushed into a hole in the road service and can be installed in minutes. These incorporate magnetic sensors which can tell when there is a car above them and can communicate this, via Telena’s proprietary ultra narrow band (UNB) network, back to a base station and thence to a cloud platform. The network operates in unlicensed 868MHz spectrum and has range and propagation characteristics that enable the Moscow city centre to be covered with 2-3 base stations. This, says Telensa, makes it cheaper than other solutions based on short range radio technologies like Zigbee; the company also cites its long track record in UNB and ability to arrange volume manufacturing as among the critical success factors.

Data created by the sensors can be used to provision road signs advising drivers where there are and aren’t parking spaces, and can also be used in conjunction with smartphone apps to both find and pay for parking. It could also be tied into information from pay-and-display payment systems so as to direct parking enforcement workers, and with tags in disabled drivers’ cars to support more effective reserved spaces.

Telensa insists that the end game for smart parking should be a fully automated parking system that supports fractional charging, dynamic pricing (something that would gladden the heart of Donald Shoup, who made a career out of arguing that curbside parking was under-priced), and more efficient enforcement. Delightfully, it is helping to drive a standardised API for smartphone parking apps, so that one day we won’t all need several different apps to park in all the places we need to.

It argues that this commercially driven approach is what’s needed to make the implementation of smarter parking stack up; the environmental benefits, and the impact on urban congestion, are not enough to sell it to most local authorities.

Telensa’s story is sort of borne out by some of the parking metrics, most of which were published by Moscow City Council, and most of which are of the ‘soft’ kind, celebrating the impact that the smarter parking solution has had on urban congestion. The various press releases put out by the city claim that traffic speeds in the city have increased by six percent (some releases claim nine percent), that road traffic is down 20 percent, that traffic density is down by 12 percent, and that the average parking time has gone down from six hours to 90 minutes. All of which is quite hard to reconcile, though the average speed data point seems to be the most tangible.

I’m not yet a complete convert to smarter parking, though I can see that an automated system has the potential to be more orderly and rational. I still think that reducing parking spaces in cities is the way to go, but a proper instrumented system can be used to support that model, and flexible variants of it, as much as a predict and provide expansion of parking.