Dept. of Transportation
Metro Transit Division

King Street Center
201 S Jackson St
Seattle, WA 98104
Metro Online Home

On-Board Automated Vehicle Location and On-Route Status

Objective 1: Technical Approach

GIS/Demonstration AVL Assessment:

Purpose:

The purpose of the GIS/Demonstration AVL assessment was to:

  • Assess the performance of the demonstration onboard AVL and on-route status function during on-and off-route coach operation;
  • Assess the potential impacts of utilizing data from a GPS based AVL system with the existing Transit Geographic Information System (GIS) map.

Caveat:

The assessment mapped demonstration onboard vehicle location data onto the Transit GIS map to analyze the relative positions of demonstration data as displayed on the GIS map. This limited analysis does not attempt to quantify the accuracy of the onboard AVL data, but provides insights into the demonstration system's logic, the KCM Onboard Systems requirements development effort, and potential Transit GIS map issues.

It should be noted that the onboard data used in this analysis did not include the final corrected GPS data available to the demonstration automated announcement and other features, and so is not representative of the actual performance of the demonstration vehicles in operation. In a full system deployment, it is expected that both the onboard data record and location data used by the onboard systems in operation would utilize enhanced GPS data.

The following summary provides highlights of the analysis conducted by MITT Geographic Information Systems staff.

Assessment 1: On-route, in service demonstration data:

GIS staff were provided with six days of data from two in-service blocks (167/11 and 7/60), taken from throughout the demonstration period. Demonstration on-board location data were mapped onto the Transit GIS map including:

  • Serviced stop events or the location where the bus opened its doors at a bus stop;
  • Unserviced stop events where the bus passed, but did not service a bus stop;
  • Distance-based events or system record of the bus location every 500 feet.
    • Two days of 167/11 data were excluded from the analysis, as the stop locations were extremely out of range, suggesting a non-onboard system related data processing error or that the system was malfunctioning.

Results:

  • Overall, when mapped on the Transit GIS map, the onboard stop event and distance-based event data were well aligned with the bus' expected route path.
  • Overall, the demonstration stop event locations coincided fairly closely with the GIS map stop locations, averaging 64 feet from the GIS stops, and ranging from 3 to 398 feet. King County Metro Transit bus zone lengths generally range from 60 to 110 feet, with 110 feet being more typical. Eighty-eight percent of the onboard stop events were less than 110 feet from the GIS stop location, or within a bus zone length.
  • The variability between onboard stop events and Transit GIS stops on 3rd Avenue in downtown Seattle, suggested the "urban canyon" effect of "raw," unenhanced GPS technology, where the global positioning satellite signal cannot be received by the vehicle if it is surrounded by tall buildings or hills.
  • The vendor AVL product provided an algorithm to correct the vehicle's determined location if the GPS signal was inadequate. In the demonstration installation, this feature was not used to correct the location data in the onboard data record, but only for the stop announcement and display functions, which worked well in downtown Seattle with no apparent urban canyon effect. The location data used in this analysis were uncorrected by the vendor's algorithm.
  • Areas outside downtown Seattle showed several instances of demonstration stop event data clustered hundreds of feet away from the GIS stop, suggesting the GIS stop location or street network may be in error. Investigation verified these stops were placed incorrectly on the GIS map by King County Metro.
Objective 4: Issues and Recommendations for Implementation

GIS impacts:

The assessment suggested areas of the King County Metro Transit GIS map where more precision is required in mapping bus stops onto the street network. The upcoming GIS transportation network upgrade is expected to improve the accuracy of the street network from approximately +/- 25 feet in incorporated King County and +/- 250 feet in unincorporated King County, to +/- 10 feet countywide. The increased accuracy is expected to contribute to more accurate placement of stops on the street network.

    • Technical Staff debriefing participants suggested that one option to explore would be to define a bus stop spatially as a zone or distance, rather than as a point, to improve the appearance of OBS stop data on the Transit GIS map, and to provide a more realistic representation of a bus stop.

Assessment 2: Off-route (out of service) data:

MITT staff conducted a limited assessment of the demonstration onboard equipment during off-route operation to assess the impact on the demonstration onboard automated vehicle location (AVL) function.

An out of service demonstration coach was logged on to a demonstration block, then operated off of its expected route at various points. GIS staff mapped the location of distance-based events, serviced and unserviced stop events, and on/off-route events recorded by the onboard system.

Results:

  • Overall, when displayed on the Transit GIS map, the onboard location data collected during off-route operation matched the actual path taken by the coach.
  • Data collected during off-route operation in the downtown Seattle Transit Tunnel suggested the ability of the demonstration equipment to accurately track the coach location degraded as the coach traveled farther into the tunnel, as GPS signals are not received underground. The Tunnel provided an example in the KC Metro operating area where a mechanism is needed for determining vehicle location when the GPS signal is not available.
  • As noted above, the vendor AVL product provided an algorithm to correct the vehicle's determined location if the GPS signal was inadequate. In the demonstration installation, this feature was not used to correct the location data in the onboard data record, but only for the stop announcement and display functions, which worked well in downtown Seattle with no apparent urban canyon effect. The location data used in this analysis were uncorrected by the vendor's algorithm.
  • The future OnBoard Systems implementation should provide a mechanism for determining vehicle location in the downtown Seattle Transit Tunnel where the GPS signal is unavailable, and provide the corrected location data to the onboard data record.

Off route status trigger:

  • The on-/off-route status indicator was triggered immediately by significant changes in the direction of coach travel, but gradual changes in direction were slow to indicate a change in on-route status. In addition, off-route operation on a street parallel and in close proximity to the coach's expected route did not trigger off-route status.
  • The Smart Bus system had many system parameters that could have been optimized to improve performance had the supplier been given an opportunity, as would be expected in a system wide implementation.
  • Data analysis indicated continuous on-route operation of the coach without servicing a stop produced an incorrect shift in the coach location data in the direction of travel. The demonstration system capitalized on frequent stops to improve the onboard AVL GPS processing. Additional evaluation and testing may be required for King County Metro service with few stops, such as express or night service.

Objective 4: Issues and Recommendations for Implementation

Technical staff debriefing:

  • Technical staff debriefing participants suggested a mechanism to trigger more frequent synchronization with GPS should be explored, such as changes in heading (direction of travel) from the gyroscope.
Updated: Sept. 2002