Seattle at a Glance

Deployment: 274 Fybr in-ground parking sensors, installed in September 2020

Location: 10-block area of Seattle's Belltown neighborhood

Curb types: Commercial vehicle and passenger loading zones

Independent field validation: 72 of 74 observed parking events detected (97%)

Measured driver outcomes: 27.9% less cruising-for-parking time and 16% less total route driving time

Research and technology ecosystem: University of Washington Urban Freight Lab, Seattle Department of Transportation, Pacific Northwest National Laboratory, U.S. Department of Energy, Fybr, Lacuna Technologies and other project partners

The deployment and subsequent driver experiment were independently documented by the University of Washington and in peer-reviewed research.

The Challenge

Finding an open curb space is a persistent challenge for urban delivery drivers. When commercial loading zones are occupied, drivers may spend additional time searching for available space, adding time and mileage to deliveries.

The University of Washington's Urban Freight Lab, working with the Seattle Department of Transportation and other project partners with funding from the U.S. Department of Energy, set out to determine whether real-time information about curb availability could make urban deliveries more efficient.

To answer that question, researchers first needed a reliable, real-time view of what was happening at the curb.

Fybr's Role

Fybr provided the underlying curb detection technology.

In September 2020, 274 in-ground Fybr sensors were installed across the 10-block Belltown study area, providing real-time, space-level occupancy data across commercial vehicle and passenger loading zones.

But the project tested more than a standalone sensor network. Fybr's occupancy data was incorporated into a broader technology ecosystem. Pacific Northwest National Laboratory (PNNL) worked with the research team on OpenPark, a parking application designed to provide real-time curb visibility, while Lacuna Technologies supported real-time data integration. OpenPark used the resulting information to show delivery drivers real-time curb availability, with changes in curb availability reflected in the application with approximately five seconds of latency.

Seattle therefore demonstrated how Fybr's real-time, space-level occupancy data could serve as an input to other platforms and applications rather than being confined to a standalone Fybr system.

Independent Field Validation

Before using the sensor data to study driver behavior, University of Washington researchers independently evaluated detection performance against recorded video.

Researchers identified 74 actual parking events during the validation exercise. Fybr's sensors detected 72 of those 74 events, approximately 97%.

The evaluation also identified 14 additional sensor triggers when a vehicle had not parked directly over the sensor. The research report discusses these observations separately from the 74 video-observed parking events.

The distinction is important because the researchers wanted the highest-resolution view of what the sensors were detecting, rather than only a filtered set of parking sessions. They were studying the sensor data itself, so retaining highly granular observations was valuable to the research.

From Real-Time Data to Driver Outcomes

The next question was more important: Could real-time curb information actually change driver behavior?

Researchers conducted a controlled experiment in which delivery drivers completed routes both with and without access to real-time curb availability through OpenPark.

The experiment involved 11 drivers completing 495 mock deliveries across 33 routes and 177 trips between July and November 2021.

When drivers had access to real-time curb availability:

Cruising-for-parking time decreased 27.9%.

Total route driving time decreased 16%.

Both effects met the study's threshold for statistical significance.

Researchers also estimated a 12.4% reduction in cruising distance, although that result did not meet the conventional threshold for statistical significance.

The results were published in the peer-reviewed journal Scientific Reports.

The U.S. Department of Transportation subsequently highlighted the Seattle findings through its ITS Deployment Evaluation program, reporting the 27.9% reduction in cruising-for-parking time and 12.4% reduction in cruising distance.

Why It Matters

Seattle demonstrated the value of treating parking detection as infrastructure for a broader technology ecosystem, rather than as an isolated application.

Fybr provided the underlying curb detection technology. That data was made available to other systems. PNNL and other project partners used the resulting information to provide real-time curb availability to drivers. Researchers then independently measured whether providing that information changed driver behavior.

It did.

Drivers with real-time curb information spent 27.9% less time searching for parking, while total route driving time decreased 16%.

For cities, the lesson extends beyond parking guidance. Reliable, space-level occupancy data can provide a foundation for curb management, loading-zone analytics, parking guidance, enforcement and third-party applications.

Accuracy in the Real World

Parking sensor performance is often summarized as a single accuracy percentage. But the value of an accuracy figure depends heavily on where it was measured, who measured it, and what was actually being measured.

Seattle provides an unusually rigorous example. The 97% detection rate was not a Fybr laboratory result, internal test or performance claim under conditions Fybr controlled. University of Washington researchers established their own ground truth using video and compared it with Fybr sensor detections during actual parking activity on public streets.

The environment matters. Commercial and passenger loading zones are particularly dynamic curb spaces, with vehicles of different sizes and positions, short dwell times, frequent arrivals and departures, and activity occurring around the parking area. Fybr detected 72 of 74 parking events independently observed by researchers under those real-world conditions.

The researchers also documented 14 additional sensor triggers that did not correspond to vehicles parking directly over the sensors. These types of brief sensor events would not ordinarily be classified as parking sessions in a standard Fybr deployment. They were retained in Seattle because the research team wanted maximum-resolution sensor data for analysis, including activity that would normally be filtered when translating sensor observations into operational parking information.

That distinction illustrates why parking detection cannot always be reduced to one number. A raw sensor trigger, a detected parking event and a parking session are not necessarily the same thing. What matters operationally is whether the system can translate activity at the curb into the information a customer actually needs.

Fybr's platform is designed around that reality. Customers can configure how parking activity is interpreted and delivered based on their particular rules, objectives and use cases. A city studying short-duration loading activity may want a different level of detail than one managing conventional parking, measuring utilization or supporting enforcement. Fybr can provide the appropriate level of information for each rather than forcing every operation into a single, rigid definition of occupancy.

That flexibility also allows Fybr to work with customers whose requirements go beyond standard configurations, adapting how parking data is delivered to support the operational outcome they are trying to achieve.

Sources

Seattle is particularly valuable as a Fybr case study because its principal findings were evaluated and published outside Fybr.

The University of Washington / U.S. Department of Energy Technology Integration Report documents the sensor deployment and independent field-validation exercise, including the 72 of 74 detected parking events and the additional sensor triggers available to researchers.

University of Washington / DOE Technology Integration Report

The peer-reviewed Scientific Reports study provides an independent link between Fybr's technology and the measured driver outcomes. The paper states that the in-ground sensors were manufactured, installed and operated by Fybr and reports the reductions measured during the OpenPark experiment.

Peer-Reviewed Scientific Reports Study

The University of Washington Urban Freight Lab's OpenPark materials identify Fybr as the provider of the 274 in-ground sensors and describe Lacuna Technologies' role supporting real-time data integration.

University of Washington Urban Freight Lab — OpenPark

The U.S. Department of Transportation ITS Deployment Evaluation subsequently highlighted the measured reduction in parking-search time and distance.

U.S. DOT ITS Deployment Evaluation

Blog

Smart Parking Best Practices for Cities: Continuous Measurement and Management

Subscribe to the Fybr Newsletter

Connect With Us