Text Size AAA
 
 
Apr 1, 2013

Smart Parking: A Launch Pad for Fully Integrated, Intelligent Transportation

 

In 2007, city drivers in a pint-sized 15-block district in Los Angeles drove more than 950,000 miles, emitted 730 metric tons of carbon dioxide, and burned 47,000 gallons of gas … searching for parking.

There’s a good bit of irony at work in those numbers, considering that there are an estimated 4.5 parking spaces per vehicle in the U.S., including 3 surface parking spaces per vehicle. Of course, don’t tell that to an urban driver searching for a parking spot in a busy city.

Some 6 years after the LA study, parking is still a major issue. A recent Freakonomics Radio podcast, “Parking is Hell,” included an interview with UCLA parking expert Donald Shoup and highlighted the high cost of free and cheap parking—economically, socially, and increasingly, environmentally.

Cities have historically addressed congestion and outsize growth by expanding off-street parking. But with dwindling land availability, strapped municipal budgets, and construction costs of $20,000–$50,000 per space for a typical downtown parking garage, it’s easy to see why cities are increasingly opting to make more efficient use of their existing parking infrastructure. Doing so also opens up opportunities to create increasingly walkable and transit-friendly urban areas as infill development can convert excess surface parking and garages to buildings and green spaces.

In particular, smart parking, in addition to promising near-term reductions in emissions and fuel use, turns out to be an ideal test bed from which to kick start a full transportation system overhaul. All the key components of an efficient, responsive, and adaptive parking system—data management, software development, innovative pricing, sensors, smartphone integration, public-private partnerships, and viable, profitable business cases—are also at the heart of a fully-integrated, multimodal transportation system. Such “Intelligent Transportation Systems (ITS)” are broadly characterized by three I’s: instrumentation (the hardware to collect and transmit mobility data), interconnection (the network to share that data), and of course intelligence (the user-facing programs and efficient processing algorithms that put the data to work to enhance personal mobility).

Data Management

Quality data is the core around which any optimized parking system is built. For the data to be usefully processed by software developers, data owners (e.g. parking structure owners, municipalities, and transit authorities) have to make it transparent, available in a standard format, and ideally centralized for one-stop access. Different data owners (municipalities, private garage owners) overseeing different functions (on- and off-street parking) tend to have their own dataset formats.

SFPark, a federally funded San Francisco parking management program, is a notable exception: one organization, SFMTA, oversees all parking and is thus well-positioned to centralize and standardize city-wide parking data. Google and others have created standard application programming interfaces (APIs) to facilitate greater coordination among developers, and software developers such as ParkMe are amassing formidable private databases by working with municipalities one-by-one to harmonize the data. Despite select bright spots, siloed data remains a persistent and costly barrier to greater adoption of Smart Parking regimes and to ITS more generally.

Software

Software is the brains behind active parking management. Real-time and predictive software algorithms process data ranging from images and time stamps at parking garage entrances to GPS coordinates on users’ phones to determine driver location and parking availability, delivering it all through a user-friendly and interactive interface. ParkMe is a parking app that uses a combination of real-time and historical data to predict parking availability for a given city block. Such real-time availability forecast (RAF) algorithms, perhaps launching from the Smart Parking test bed, could be central to better traffic management since accurately predicting jams and roadblocks is critical to providing users with alternatives that can avoid or overcome them.

Visualization and integration with existing web applications could be key to scaling transportation and parking software solutions. Parker, an app provided by Streetline, recently released a custom map generator that allows merchants, universities, cities, and parking providers to showcase nearby parking options via any website.

Sensors

Steetline, SFPark, and LAExpressPark, among others, rely on physical sensors embedded in the pavement in each parking spot to generate parking data. Most sensors detect something in their proximity through ultrasonic technology. They sell the sensors to their customers, primarily municipalities, and usually provide a software app along with them as one package.

Smartphone Integration

Other companies, ParkMe among them, wonder whether quality data can be generated without the sensors, whose installation tends to be a large expense—about $1400 each according to the Transportation Sustainability Research Center at UC Berkeley.

Smartphones are now equipped with GPS with resolution around ~10m and they continue to approach the quality of handheld GPS units with resolution within 1m—good enough to differentiate among individual parking spots. Another smartphone-based approach is to use a “Hybrid Indoor Positioning Engine” (HIPE), which fuses Wireless Local Area Network signals (currently used in the “Current Location” function of many map apps) and the measurements of the built-in sensors of smartphones to determine both location and mode of transport.

A smartphone application called on{X} provides an open-source platform allowing users to program their phones to automatically sense when they are parked. The application works by actively monitoring the phone’s built-in accelerometer and sensing a change from driving to walking. It then automatically records the location of your parking spot through one of the locating techniques described above.

Smartphones and connected vehicles with embedded software capabilities can also perform real-time parking space selection optimization, in part based on real-time awareness of other vehicles’ location, to revise a previous parking spot choice in case a previously-selected space is taken by another driver or a closer spot becomes available. Users provide parameters such as proximity to destination and parking cost, and the app does the rest.

Payment can also be seamlessly completed through the phone with apps like PayByPhone.

Innovative Pricing

Shoup of UCLA and the Freakonomics Radio broadcast has estimated that 30 percent of city drivers in dense urban areas are searching for parking. By employing demand-responsive pricing, parking can be shifted away from peak times and areas, increasing availability and throughput (and therefore parking revenue) while reducing traffic congestion. With dwindling revenues from gas taxes looming as vehicles become more efficient, such pricing mechanisms can help to bolster budgets for infrastructure improvements even as they provide the data indicating where improvements are most needed.

Public Private Partnerships

Because private entities tend to provide the software and public entities control and generate the data, public-private partnerships are critical to good parking management. Streetline has an impressive ecosystem of partners to develop everything from sensors to cloud computing to municipal financing options through a partnership with Citi. Effective collaboration will be key to fusing a data landscape currently riven by myriad approaches, formats, programs, and jurisdictions.

Proving out the Business Case

Benefits of smart parking for municipalities include: fewer attendants, better collection, higher throughput, better utilization, and reduced emissions. User benefits include fuel and time savings, reduced hassle, and more flexibility (e.g. the freedom to stay for dessert by extending parking remotely). Businesses potentially benefit from more customers and the opportunity to work with software developers to engage in targeted (location and demographic-specific) advertising for their services.

Of course, these systems aren’t free. Costs include sensors (if employed), data routers, and labor associated with organizing and making data available.

Listing the elements of cost and value is easier than quantifying them, especially for parking, but a few studies do exist for ITS systems in general: a study examining ITS deployment in Beijing estimated an investment between 2005 and 2008 of 1.2 billion Yuan while the system’s social and economic payoffs added up to 26.8 billion Yuan, a return on investment of 22 times. The State of Michigan completed a similar assessment, concluding that ITS is responsible for a savings of $134 million USD annually.

Descending the Learning Curve

Many challenges remain, of course. Data standardization is a persistent challenge, sensors are not 100% reliable, and data transfer delays lead to parking availability inaccuracies. Then there’s the issue of “placard abuse,” discussed in the Freakonomics segment. In Shoup’s backyard, the City of Los Angeles, up to 40% of parkers legally avoid paying for parking spaces, largely through the liberal distribution and use of handicap credentials. Such placard abuse drives twice as much parking revenue loss as negligent payers and poses a potentially large hurdle for smart parking systems that employ forms of congestion or demand-response pricing, since placard abusers are insensitive to price fluctuations and exact a growing financial toll on the system through lost revenue during high demand periods. (A “digital placard” may have a role to play here.)

Never the less, the microcosm of smart parking provides a rare natural laboratory to test, multiply, and scale the solutions for an instrumented, interconnected, and intelligent transportation system as a whole. Despite some challenges, its opportunities—data integration and standardization, innovative software to process and deliver the data, sensing technology, better municipal infrastructure planning with revised revenue streams, and business models built around empowering individual users through smartphones and connected vehicles—are the nodes from which a transformed transportation network can ultimately emerge.

Recommended Reading 

First image: Kay Dollfus / Shutterstock.com
Second image:
Suzanne Tucker / Shutterstock.com
Third image; Courtesy of
Shutterstock.com

========

This blog post also appeared on Greenbiz.

Join the Discussion


Showing 1-2 of 2 comments

April 4, 2013

And -- Another thing --
Some years ago I was taking a Continuing Education Class at University of Colorado Denver. It was in the evening. The Denver campus is fairly busy in the evening. Parking can be difficult. There was, however, a fairly large parking lot that was permitted by "Permit Only." Set aside for faculty and other University personnel.
Our instructor said he didn't see any reason we could not park there because most of the faculty etc. was gone in the evening. Ho Ho! I got a ticket anyway - from the University police. No exceptions.
There was a large vacant parking area in a high demand area - but off limits to most drivers.
There are many cases like this.
Does your Smart Park System have any thoughts about this?


April 30, 2013

It doesn't seem to be a good idea to make parking/road info too dependent on SmartPhones (especially as occupancy sensors). Freiburg, Germany has signs along the roads entering the city center showing available parking spaces in each lot (I think with mechanical indicators)-- much better than serving only phone users who might make illegal accesses while driving. Road signs serve all drivers conveniently.

Most parking lot "full" sensors count cars entering and leaving a lot, so don't try to sense each space. But Sunnyvale, CA is replacing non functioning magnetic traffic sensors with cameras that count cars by measuring a difference against the background pavement-- I think they are much cheaper. Parking lot occupancy could be measured by a grid of $10 webcams+$5 microcontroller+$5 Zigbee (or other RF mesh) network sensors-- I don't know if anyone sells this. Placed on a light pole, the sensor might process an image like the one shown above. An upgrade could also be used as security cameras.

A standard open-source API would be key to interoperability among parking lot or street sensor manufacturers, and the various display alternatives and analysis software. I don't mean an end-user map API for one proprietary web site-- I mean a way for various parking lot sensor systems to exchange data with the public-- road signs as well as various web apps and analysis sites. This might mean an API for sensors to store data in a server, plus an API to query sensor status.

PAGE: 1