Transportation Techies: Safer Streets with Better Data

Here is my first post as Mobility Lab’s Tech Reporter, originally published here.

Gallo, MVJ web

Washington, D.C.-area traffic safety has improved significantly over the past few years, but there is still room for more progress.

In the most recent installment of the Mobility Lab-sponsored group Transportation Techies, Playing with Traffic, presenters offered their insights for making our streets safer and more efficient for people.

According to new data from the Metropolitan Police Department, more pedestrians die in crashes than people in cars. From 2009 to 2014, 58 pedestrians were killed from vehicular crashes, compared to 51 drivers or passengers. Nationally, while overall traffic deaths fell in 2014, pedestrian deaths rose by 3.1 percent.

Jacob Mason, of All Walks DC and the Institute for Transportation & Development Policy, provided important context for what it would take to truly address pedestrian deaths. Open transportation data is immensely important for residents to “more easily participate in government” and, regarding transportation, “identify traffic patterns and safety problems.” For pedestrian crashes, in particular, Mason noted that it’s easier to find detailed information on the District’s roadkill than its crashes – and especially areas so dangerous most pedestrians won’t even attempt to walk.

Understanding the need for better data, two major themes emerged from the evening’s presentations. One was the pursuit of improving the efficiency of roads and the challenges of using models to do so. Second, a daily effort for drivers, was the effort to increase driver accountability and make them safer road users.

Planning from above

Will Kasper, with Skycomp, shared how his company uses aerial photography to understand and address traffic problems. “All models are wrong, but many are useful,” Kasper quoted, noting that many models use computer simulations without actually having a real system to understand first. Skycomp’s approach, on the other hand, uses real-world data input to inform traffic studies and modeling.

Skycomp, MVJ

By tagging and following thousands of individual vehicles throughout an established observation area, Skycomp creates a time-lapse that helps to visually understand traffic problems. The company’s sensors collect immense amounts of data, such as travel times and even driver behavior, and determine lane-by-lane information, rather than just aggregates across the route. This helps provide more accurate models for planners, and they can go back to filter routes with incredible granularity in order to fine-tune the process.

On the application side, Anthony Gallo of Kimley-Horn is working on traffic simulation for the I-66 Outside the Beltway Project. Reflecting Kaspar’s approach, Gallo examined the use of modeling to create traffic plans, and reconciled the differences between virtual simulations with actual data from the project area.

Thus, the project has taken over a year to calibrate and accurately represent the frustrations of rush-hour commuting on I-66. According to Gallo, a major challenge for such a project is coming up with a quantifiable justification for why a funding-intensive “build scenario” is more effective than a free “do-nothing scenario.”

Through his model, Gallo acknowledged shortcomings, such as the inevitability of induced demand following the increased lane capacity. Traffic is just about unavoidable in the general-capacity lanes, but High Occupancy Toll lanes would provide the option of a free-flowing facility for transit users and drivers who choose to pay for the convenience. Gallo emphasized that this is a major intention of the project, and that the approach would offer an important benefit in choice by going multimodal to avoid congestion.

In all, his modeling has shown that the project will not solve congestion, but would at least alleviate it by providing more options to commuters.

Less texting, safer driving

Distracted driving is a massive safety problem, accounting for about 10 percent of traffic fatalities every year. But a major contributor to this problem – mobile devices – can actually be put to good use to increase driver accountability, and ideally their attention to the road.

Jared Sheehan, Android lead with Driversiti, demonstrated the Driversiti app, which uses phone sensors to detect driving performance and conditions, and then analyzes specific events to create a black box-like record of crashes or other behavioral anomalies.

The easiest example of this is that the phone could tell if there has been a crash. Driversiti has crashed a number of test cars with phones placed throughout the vehicles, to optimize the sensors’ understanding of crash conditions.

It also maintains a database of road conditions. For example, if the phone senses that multiple cars have slipped on black ice, it can broadcast the hazard to other users and suggest rerouting. Sheehan described this as “situational awareness.”

Fleet and insurance companies find this technology especially appealing, Sheehan noted. Because the app tracks telematics – speed, acceleration, deceleration, swerving, and so on – companies can use the software development kit to sense unsafe nearby drivers. The sensors can also detect distracted driving, and hold drivers accountable either through insurance rates or company intervention via Driversiti’s app.

On a more community-based angle, Brian Zito, Ashley Peter and Seth Clark of Booz Allen Hamilton are developing an app named “Eyes Ahead” to discourage texting and driving through group observation and awareness. Eyes Ahead hopes to create a communal support system to mutually encourage users to become safer drivers.

Eyes ahead, MVJ

The program, which runs in the background while a user drives, pulls texting data and records the speed of the car at that time. This information is then recorded on a group website, creating a group incentive: distracted drivers would have to explain themselves to their community if the program shows them texting and driving.

It’s still a work in progress, but hopefully the app will provide a new tool for family and friends to become involved in improving each others’ habits.

Rakesh Nune from the District Department of Transportation presented a project that uses slightly different data from phones in traffic: social media. He has built an application that relies on Twitter to speed up operator response to crashes. Social media, he posited, helps bridge the coordination gap from incidents to emergency responders and then to DDOT for cleanup. Tweets provide DDOT with contextual details about each crash that the agency could not otherwise glean from other traffic applications.

By monitoring Twitter for keywords related to crashes, such as “crash” or “#dctraffic, his program aggregates a mass of informative – if informal – traffic reports. The tweets matching keyword combinations are organized into official, news, and citizen tweets to prioritize and verify where the District needs to direct its resources. Combined with Waze, a community-based traffic and navigation app, DDOT can glean both the geographical layout and a contextual sense of what happened so it can analyze the data, if needed, later on.

And don’t worry, DDOT isn’t actively encouraging anybody to tweet while driving. The project has no outreach component, and the majority of informational tweets actually come from (non-driving) news sources.

For more information on Transportation Techies, join the Meetup group. The next meetup, which will focus on bike-related projects, is December 16 at 1776 Crystal City.

All photos by M.V. Jantzen, Flickr. Top, Anthony Gallo presents his models of expansion options for I-66.

– See more at: http://mobilitylab.org/2015/12/07/safer-streets-with-better-data/#sthash.nHGJif0S.dpuf

Advertisements

1 thought on “Transportation Techies: Safer Streets with Better Data”

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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