Assure accurate late train arrival forecasts in support of a Connection Protection system.
Experience of the Utah Transit Authority in implementing a Connection Protection program for rail-to-bus passenger transfers in Salt Lake City.
Made Public Date
04/13/2006

657

Utah
United States
TwitterLinkedInFacebook
Identifier
2006-00227

Evaluation of Utah Transit Authority's Connection Protection System

Background

The Utah Transit Authority (UTA) implemented a Connection Protection (CP) system to improve the reliability of transfers from the higher frequency light rail trains to the lower frequency bus services at selected connecting rail stations. The CP system examines the status of trains and issues a "hold at {station name} until {time}" message to buses waiting at the connecting rail stations via the bus' onboard Mobile Data Terminal (MDT), if the lateness of train is within a pre-determined time threshold. The system was completed and tested in January 2002 prior to the Winter Olympic Games in Salt Lake City. Operationally, the CP messages are triggered by the late train events. One late train could potentially trigger multiple CP messages to hold one or more buses at each of the next three down-line stations, based on the CP assignment. An evaluation was conducted to assess the performance of the CP system and the response of passengers and system operators. This lesson is based on experience with the early operation of the CP system with the objective of identifying potential improvements and offering guidance to other transit agencies elsewhere that may be considering implementing a similar system.

Lessons Learned

The CP system uses an algorithm to predict whether trains will arrive late at a station and whether a "hold until" message should be issued to buses at the station. Essentially, the system monitors train arrival times at selected points along the rail line, determines if the train is running late, assesses how late the train is expected to be, compares this projection against the scheduled bus departure times, and issues a CP "hold until" message to the appropriate bus trips, at up to three downstream rail stations. A significant number of CP "hold" messages were issued by UTA to bus operators where the train was actually less than two minutes late (38%), indicating that trains that were projected to be two or more minutes late were eventually able to make up time after the point from which the initial projection was made – resulting in inaccurate issuance of CP "hold" messages. Once a CP message is sent, no updates or recall is issued when the train's status changes from the initial projection. In addition, the evaluation pointed out that trains can also be delayed more than projected, resulting in a failure to issue a CP "hold until" message when appropriate. Inaccurately issued CP messages can have a detrimental impact on bus operators' motivation to comply with CP. On the other hand, if CP "hold" messages are not issued when they should be, it will be less effective than it could be because bus operators will not receive guidance from the CP system.

Some guidance from UTA's experience with implementing their Connection Protection system, specifically related to train arrival estimates and forecasts, include the following:

  • For CP systems based on a train arrival forecast, seek to update train status and down stream forecasts frequently. This will complicate the process by requiring more frequent status checks, projections of revised arrival times, and notifications to bus operators. However, a presumption is that operators will be more likely to buy in and adhere to CP directives when "false" train delay messages are reduced through more frequent updating.
  • For systems in which GPS-based train location information is available, feed accurate train location data into the CP algorithm in real-time on a frequent basis. Eventually, when Automatic Vehicle Location (AVL) systems are available on all the buses and trains, and location data shared in real-time, the CP algorithm can be applied with much greater precision, resulting in substantially improved performance and positive operator and customer satisfaction.
  • Use historical performance data, where appropriate, to fine tune train arrival forecasts. These data might include accurate measures of variation in travel times between stations, loading time at each station by time of day and level of demand, and the ability of trains to make up late time in traveling between stations.
  • Notify bus operators as soon as revised information contradicts an earlier issued CP "hold" message. While this would result in additional complexity to the system as noted above, it would be helpful to both the bus operators and the on-board passengers to receive notification either that required wait time has been reduced or that waiting is no longer required.

Efficient and accurate issuance of CP messages is expected to enhance customer satisfaction and encourage higher CP compliance by bus operators. A well-designed CP system can help achieve ITS goals to enhance efficiency, mobility, productivity and customer satisfaction. The application of a Connection Protection program in Utah has served as a useful tool that can help operators better meet the needs of their transit customers. The evaluation of this system identified the importance of accurate knowledge of both train and bus locations, and incorporating this information on a real-time basis into the CP algorithm.

Evaluation of Utah Transit Authority's Connection Protection System

Evaluation of Utah Transit Authority's Connection Protection System
Publication Sort Date
05/12/2004
Author
Jeffery Jenq, Chris Cluett, Ben Pierce and Alan Pate, Battelle
Publisher
ITS Joint Program Office, U.S. Department of Transportation

(Our website has many links to other organizations. While we offer these electronic linkages for your convenience in accessing transportation-related information, please be aware that when you exit our website, the privacy and accessibility policies stated on our website may not be the same as that on other websites.)

Focus Areas Taxonomy: