2021 07 28 Worker Presence Subgroup meeting - usdot-jpo-ode/wzdx GitHub Wiki

Meeting Information

July 28, 2021, 1:00-2:00pm EDT

Purpose

  • Review and discuss open GitHub issues (#39, #173, #182) related to worker presence.
  • Present and discuss proposed structure of WZDx specification update.
  • Highlight upcoming outreach activities.

Agenda

  • Welcome and Sign-in
  • Worker vs. Equipment Tracking
  • Proposed Update to WZDx Specification
  • Next Steps and Action Items

Minutes

Workers vs. Equipment Tracking - #173 (Serge)

  • In some areas of the country, the worker presence concept includes many scenarios. In others, it's really just about workers in or near the roadway.
  • Right now, the spec has a Boolean (true/false property) of whether workers are present.
  • Some folks would rather focus on tracking equipment since that’s generally easier. We want to focus on the measurement of real people, but tip our caps equipment-based data that can be used to infer worker presence.
    • For equipment itself – leave for a future extension or through the work of the SWZ Subgroup.
    • Right now the WP co-chairs see them as two different things
  • To account for different definitions in different jurisdictions, we're proposing is to have a list of choices for defining worker presence - this slide has some potentials

Proposed Spec. Update - #182 (Kristin)

The goal of this subgroup is for all producers to include WZDx in their feeds

  • How do we get more data providers to share worker presence info?
  • Current version of how worker presence is expressed is not nuanced enough for all data producers to feel comfortable including the info

Current workers_present property: Boolean indicating that there are workers present in the road event

  • Yes or no doesn’t work great here because worker presence has different meanings in different jurisdictions

Option 1 - At individual work zone level, flesh out the spec with some more fields to indicate more about how workers are present.

  • Are workers present: true/false/unknown?
  • Source:
    • Serge talked about this in his slides - don't just say that workers are present, but say how we know
    • Enumerated values of how we know there's someone there
  • Last confirmed:
    • Sharing when worker presence was last verified
  • Reliability score: 8
    • Data producer can provide a reliability score based on how confident they are in their data
    • Inspired by Waze's reliability scores on a scale of 0-10. they have a secret sauce for coming up with reliability, but this would be up to the agencies
  • Definition:
    • In some places there are different definitions of worker presence, and there are some legal ramifications for what that definition is

Option 2 – include jurisdiction’s worker presence definition in the metadata (RoadEventFeedDataSource) to avoid redundancy at the RoadEvent level

Build your WP definition - spec would have a bunch of descriptions about what worker presence could be

  • Looking for feedback on the building blocks
  • We have a list of different situations, and a producer can build out what worker presence means for them (legally or otherwise)

Source method - also ties in with what Serge presented on

  • Want to communicate how we know that workers are present
  • Also want more feedback on the options

Are_workers_present – should we add unknown to yes/no? Absence of a worker presence object means "I don't know" Other way forward is to allow users to say I don't know - could advance requiring the field

Discussion Summary

Dan Sprengeler: What does active mean in event status? If a work zone is inactive, that to me means that signs are up but work isn't ongoing. I'm more interested in are_workers_present incidating whether workers are vulnerable

  • Jacob: event_status is somewhat redundant with start_date and end_date but that’s kind of a can of worms. In the case where start and end times are long and span multiple days, say 'active' when there is traffic control in place.
  • Ross: When you start talking about status and worker presence, you get into a messy area. Worker presence might be part of status. It might be appropriate for us to agree to address it separately. Jurisdiction to jurisdiction, these fields' relationships might vary.
  • Serge: The Spec. Update and Worker Presence groups will chat to fill the gap between those fields.
  • Ross: The publisher of project needs to be the authority on this information.
  • Dan: We have to walk before we run. The data producer is feeding the field. If in the future we need to relocate, work on it then.

Chris: Michigan DOT still doesn’t know whether workers are present most of the time. And then you wonder how long you tell someone is active for. If someone is present within a short segment of a 20 mile work zone, are we going to say workers are present in the whole work zone? Or are we going to have separate events for where workers are actually working?

  • Chris: I would prefer yes/no/unknown for are_workers_present. It allows the specification to have a placeholder so that when we know more information it's expanded upon.
  • Chris: Consider having an event for a worker. Drivers will ignore warnings if they haven't seen anyone.
  • Kristin: This is different for different agencies. Some are more comfortable with providing worker presence. They're optional fields and don't have to be reported, but we should give space to evolve it over time.
  • Jacob: The worker could be its own GeoJSON feature, with its own geometry, separate but linked to a road event.
  • Serge: That separate event could highlight a region, or specific worker. A lot of states don't have all that information, so this is still an optional field.

Kristin: We’ve heard that there is a need for worker presence data - can any data consumers on the call expound on that?

  • Eli: GEWI is more of an aggregator for end users. From my experience, knowing a worker is there is less important unless they're in a travel lane. For automated vehicles, we want standard data about lane closures. If workers are moving across the lane, then AVs will look to see whether someone is across the road. If behind barriers, it shouldn't be as much of a factor – someone behind a guardrail isn't going to affect a driver’s behavior.
  • Ross: But in multiple jurisdictions a worker being present (even behind guardrails) changes the law. Fines are different if they are in the work zone vs. not.
  • Eli: Speed changes are something drivers need to know. Is that going to be the case if closure is only in a travel lane? Trigger of worker presence doesn't mean anything to driver
  • Ross: Do we take this from the point of view of the authority, or from the vehicle operator?
  • Eli: In the event of a collision, does a WZDx feed become evidence in a legal proceeding? Who filled out that field?
  • Kristin: We're at risk of analysis paralysis here. We want to add a flexible forum to take baby steps.
  • Eli: How useful is it? DOT offices sometimes don't always have consistency about what data is provided. We're happy to use it and make it available to our consumers.
  • Dan: It's not just about vehicles on highway interacting with traffic. If you want to reduce worker injury, focus on their own equipment.

Next Steps

  1. Worker Presence co-chairs will initiate creation of a pull request implementing the proposed changes
  2. Subgroup members should comment on the GitHub issue and (when created) pull request
  3. Co-chairs are also planning outreach to data consumers to review the proposed change to the specification and other worker presence data needs

Participants

Name Organization
Kristin Virshbo* Castle Rock Associates
Serge Beaudry* Ver-Mac
Luke Urie* Austin Transportation Department
Nagham Matout ATSSA
Mahsa Ettefagh Booz Allen Hamilton
Eli Sherer GEWI NA
Curtis Hay GM
Todd Hartnett Hill and Smith
Fabio Capillo Houston Public Works
Michelle Boucher IBI Group
Jacob Brady IBI Group
Ross Sheckler iCone
Dan Sprengeler Iowa DOT
Faisail Saleen Maricopa County DOT
Hua Xiang Maryland DOT
Neil Boudreau Massachusetts DOT
Chris Brookes Michigan DOT
Michelle Moser Minnesota DOT
Ted Ulven Minnesota DOT
Eneliko Mulokozi RTC of Southern Nevada
Martha Kapitanov USDOT Federal Highway Administration
Todd Peterson USDOT Federal Highway Administration
Molly Behan USDOT Volpe Center
Nate Deshmukh Towery USDOT Volpe Center
Mark Mockett USDOT Volpe Center
Hadrian Merced Hernandez USDOT Volpe Center
Tony Leingang Washington State DOT

* Co-chair of the Worker Presence Subgroup