Replace beginning_accuracy and ending_accuracy with postional_accuracy.

[DeraldDudley]

Replace beginning_accuracy and ending_accuracy with postional_accuracy.

The properties beginning_accuracy and ending_accuracy were established in version 1 of the specification before the adoption of GeoJSON. They were important because it was thought a work zone would be located solely by a beginning point and and ending point. The beginning_accuracy and ending_accuracy properties were essentially feature level metadata records established to build confidence in the spatial accuracy of the data.

With the adoption of GeoJson, features needs only one property to describe their accuracy. The industry standard describes a feature's positional accuracy. Here is the FGDC Accuracy Standard. This is a nice Positional Accuracy resource from University of Minnesota.

Proposed Changes...

1. In the RoadEvent feature;

   • Drop beginning_accuracy
   • Drop ending_accuracy
   • Insert positional_accuracy
   • Value of positional_accuracy is the PositionalAccuracy Object.

2. Change Spatial Verification enumeration to Positional Accuracy Object

Positional Accuracy Object

The PositionalAccuracy object indicates the spatial accuracy of the feature.

Properties

Property Name	Data Type	Description	Conformance	Notes
verification	enumeration: estimated or verified	An indication of how the feature was defined.	Required
accuracy	real	A measure of the spatial accuracy	Optional	This property sets the specification up nicely for measured accuracy.
UnitOfMeasurement	enumeration: string	Unit the accuracy value represents	Optional	This property sets the specification up nicely for measured accuracy.

Used By

Property	Object
positional_accuracy	[RoadEvent](/spec-content/objects/RoadEve

[sknick-iastate]

@DeraldDudley I'll need to dig into the accuracy standard but it appears to apply to point features not linear features.

Objective: The National Standard for Spatial Data Accuracy (NSSDA) implements a statistical and testing methodology for estimating the positional accuracy of points on maps and in digital geospatial data, with respect to georeferenced ground positions of higher accuracy

Again, I'll try to dig into it more but not sure how applicable it is to linear features. I still feel like there is a need to convey how accurate the begin and end location of the work zone is. In Iowa we will be verifying the begin of the work zone with a smart arrow board but the end of the work zone will still be an estimate.

[DeraldDudley]

@sknick-iastate - Because lines and polygons are made of points its good for all geometry types.

If we're tracking the accuracy of individual points, and only using two points to define the location of a work zone, should we consider defining the work zone with two separate, but related, features.

From the U Minesota Report - "The National Standard for Spatial Data Accuracy describes a way to measure and report positional accuracy of features found within a geographic data set. Approved in 1998, the NSSDA recognizes the growing need for digital spatial data and provides a common language for reporting accuracy."

[j-d-b]

Some negatives I see:

• Loses the ability to indicate that the start is verified but the end (or rest of linestring) is not (or any combination of mixed verified/estimated points)
• Adds complexity (new object, more nesting, more to learn, more processing logic/code)
• There has not yet been desire from producers or consumers for a precise measurement of accuracy
• Equipment providers have indicated that even if the device reports down to the centimeter there is GPS float and similar that make it so the "accuracy" is hard to determine. Based on the last spec update co-chairs meeting, it would be better to just indicate in the spec for a position to be "verified" it has to meet a defined threshold (e.g. 1 meter) of accuracy.

[DeraldDudley]

This thought came to mind when I was looking at the object model and trying to figure out how inheritance should work for a "transportation data exchange feed". Typically, feature level metadata, like positional accuracy, describes the whole feature; Not individual parts of a feature. Seems like describing the positional accuracy of only the first and last point of a feature is unique to the work-zone EventType.

It also occurs to me that we should be concerned about the positional accuracy of the geometry in between the fist and last points of a work zone. How do we describe that?

Draft RoadEvent Inheritance Model

[DeraldDudley]

After today's extension meeting I'm providing a nested model...

Draft RoadEvent Nesting Model

[DeraldDudley]

@j-d-b Can we close this one?

[j-d-b]

Closed following Spec Update Subgroup meeting on 2022-02-23.