Click for next page ( 12


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 11
11 software logs would correspond to one inferred from an O-D 2004 and 1995, with the latter being the year covered by the survey. Likewise, "inferences" from customer complaints, Volpe Center's first deployment survey. For AVL, the num- which often refer to a person's unpleasant encounter with an ber of properties with operational systems grew from 22 in operator, are unlikely to correspond to riders' assessment of 1995 to 157 in 2004, a 614% increase. Adding planned de- their treatment by operators from a customer satisfaction ployments, the number of AVL properties grew from 86 in survey. 1995 to 257 in 2004, an increase of 199%. Under the limiting Subsequent sections of this report include the use of ITS data assumption that none of the non-responding properties had from customer-initiated contacts. It should be noted that ap- or were planning to deploy an AVL system, the correspon- plications drawing on such data mainly support customer serv- ding industry penetration rates for this technology in 2004 ice and human resource rather than market research functions. were 30.4% (operational) and 49.8% (operational plus planned). The deployment of APC systems in the transit industry has Inventory of ITS Data been more limited than AVL deployment. Only 11 properties for Market Research reported operational APC systems in 1995, and the growth to 75 properties in 2004 resulted in an industry penetration rate Deployment just under half the corresponding AVL rate. Adding planned The transit industry's use of ITS data for market research is APC deployments yields totals of 32 properties in 1995 and conditioned by the extent of deployment of the respective 129 properties in 2004, with industry penetration of 25%. technologies. The deployment status of advanced public Departing from previous surveys, the Volpe Center did not transportation technologies has been systematically tracked report fleet coverage information for vehicle-related tech- over time by the U.S. DOT's Volpe National Transportation nologies in its 2005 report. Earlier reports indicate that AVL Systems Center. The Volpe Center's most recent report docu- and fare payment technologies are commonly deployed ments ITS deployment in 2004 (Volpe Center 2005). Data on fleetwide, while APC systems are not. A general "rule-of- ITS deployment status in 2004 were recovered by a survey of thumb" in the industry is that 15% fleet coverage is the 516 transit properties that report information to the NTD. minimum necessary for APCs to satisfy NTD reporting There were 327 responses to the survey. While the deployment requirements. Anecdotal evidence suggests that properties status of advanced technologies among non-respondents with APCs are increasingly deploying the systems well beyond is unknown, it is reasonable to assume that properties with the rule-of-thumb threshold in order to support internal operational or planned technologies were more likely to reporting and analysis needs. respond to the Volpe Center survey than those without the Coverage of electronic fare payment technologies in the technologies. Volpe Center report is limited to magnetic stripe and smart Deployment information from the 2005 Volpe Center re- card systems. In 2004, there were 159 properties with opera- port is presented in Table 2-1 for the technologies of interest tional card systems, up from 22 properties in 1995. Adding to this Guidebook. Information in the table is presented for planned systems increases the respective totals to 65 and 293. Table 2-1. ITS deployment in the transit industry, 19952004. Electronic Automated AVL APC Fare P'mt Transit Info. 1995 Operational 22 11 22 48 Operational + Planned 86 32 65 93 2004 Operational 157 75 159 488 Operational + Planned 257 129 293 512 Percentage Change, 1995-2004 Operational 614% 582% 623% 917% Operational + Planned 199% 303% 351% 451% Industry Penetration, 2004 Operational 30.4% 14.5% 30.8% 94.6% Operational + Planned 49.8% 25.0% 56.8% 99.2% Properties Surveyed, 2004 516 516 516 516

OCR for page 11
12 The industry penetration rates for electronic fare payment become increasingly favored over magnetic stripe cards. systems are comparable to the rates for AVL systems. Especially when registered, smart cards are capable of re- Automated transit information technologies are defined as covering more information about riders and their use of ". . . systems that provide information to the public, without the system. human intervention . . ." (Volpe Center 2005: 122). The The deployment of APCs has not been as extensive as the media through which information is provided include the other technologies covered in Table 2-1, yet the passenger Web, automated telephone systems, automated voice an- counts provided by these systems hold great potential for nunciation systems (AVA), signboards, cell phones, pagers, leveraging traditional market research applications. and personal digital assistants. The types of information pro- vided include schedules and fares, trip plans, and real time vehicle arrival times. The deployment data on this collection ITS Data Inventory of technologies indicate that survey staff was able to recover information about virtually all non-responding properties, The ITS technologies recovering data for market research most likely by visiting the properties' websites and calling can be grouped into two categories. The first includes systems their phone systems. In 2004, 488 operational and 512 oper- that are installed on vehicles, including AVL, AVM, MDTs, ational plus planned systems were identified, compared to 48 and control heads, APC, magnetic and smart card readers, and and 93, respectively, in 1995. The 2004 figures indicated that electronic registering fareboxes. The second category includes the penetration level in the industry is near-complete. systems supporting customer service, including ticket vending In addition to the technologies listed in Table 2-1, the machines, automated phone systems, and the Web. Traffic Volpe Center began reporting of the deployment of mobile counting systems are also included in the second category. data terminals in its 2002 survey. When installed on transit Data elements recovered from on-board systems are pre- vehicles and integrated with AVL systems, mobile data ter- sented in Figure 2-3. The role of AVL systems in recording minals can record pre-programmed events when operators time and location information is central to the viability of press a key. There are a number of conceivable events whose other on-board systems, in addition to its independent use- documentation would provide useful information to market fulness. When integrated with AVL, the data recorded by the researchers. The survey reported 115 transit properties with other on-board systems become identified with respect to operational mobile data terminals and 137 additional prop- where and when a specific event occurred. With systems in- erties with planned deployments in 2004. This corresponds tegration, the records generated through AVL also serve as with industry penetration estimates of 22.3% (operational) the basis for aggregating other on-board data, from unique and 48.8% (operational plus planned). events that are recorded at specific locations and times, to There are a number of general conclusions that can be summaries at higher levels, such as route and system (spa- drawn from the information in Table 2-1 and from closer in- tially) and hour and month (temporally). spection of property-level data provided in the series of Volpe The integration of AVL with other on-board technologies Center deployment reports: has been problematic, especially for properties that acquired on-board systems over time in a piecemeal approach (Casey The pace of technology deployment over the 10-year pe- 2000). Casey pointed out that the complex proprietary soft- riod has been rapid, suggesting parallel transformations in ware provided with the early systems hampered integration data archiving infrastructure, data reporting and analysis efforts, as did each property's desire to customize a given sys- practices, staffing needs, and skill requirements. tem to meet its specific needs. Integration problems have ITS deployment was initially concentrated among larger been serious enough in some cases to require installing properties and later spread to smaller properties. Current multiple AVL units on a vehicle, with each dedicated to pro- deployment levels are still higher among larger properties. viding time-location referencing to one or several on-board Two implications follow from these observations: systems. (1) there has been an opportunity for smaller properties to AVL integration with other on-board systems has im- learn from the ITS deployment experiences of larger prop- proved, following the development of the "smart bus" con- erties, through peer exchange and other means of commu- cept (Furth et al. 2006). The concept is organized around use nication and (2) if industry penetration measures were of a vehicle logic unit for storing data recovered from on- based on the number of customers affected rather than on board systems. On-board systems are now manufactured to properties, the extent and impact of deployment would be comply with the J1708 integration standards used in most much greater than indicated in Table 2-1. AVL systems (Society of Automotive Engineers 1993, 1996). Among electronic fare technology deployments and Focusing on AVL independent of the other systems, the planned deployments over the decade, smart cards have most useful data for market researchers are the time-location

OCR for page 11
13 Automatic Vehicle Location 1. Date & time 2. Vehicle ID 3. Route ID 4. Trip ID 5. Operator ID 6. Location (latitude & longitude, Stop ID) 7. Direction 8. Arrive Time 9. Depart Time 10. Speed (average or maximum, from previous location) Automatic Vehicle Monitoring 1. Door open & close 2. Lift deployment 3. Signal priority request Mobile Data Terminal/Vehicle Control Head 1. Event Code (fare evasion, pass up/overload, security issue, etc.) Automatic Passenger Counter 1. Boardings 2. Alightings 3. Load Electronic Farebox 1. Transaction ID (if transaction-based) 2. Transaction Type (ticket/token/cash) 3. Transaction Amount 4. Ridership (operator initiated) 5. Event (operator initiated, e.g., fare evasion, lift use, pass use, transfer use, etc.) Magnetic Stripe Card (Reader) 1. Card ID 2. Card affiliation (when registered) 3. Transaction ID 4. Transaction Amount Smart Card (Reader) 1. Card ID 2. Person ID (when registered) 3. Transaction ID 4. Transaction Amount Figure 2-3. Inventory of ITS data for transit market research: on-vehicle systems. vehicle status data. At the lowest level of aggregation these a deterioration in regularity results in "bus bunching," which data are collected at stops and route origins and destinations. leads to longer waits for riders, a higher incidence of crowd- Aggregating the time data over all route locations yields a ing on vehicles, and a loss of effective service capacity for the vehicle's actual running time for a given trip. Analysis of the transit agency. The ability to know where and when such pattern of running times for all trips for specified time periods problems occur facilitates the operations control process reveals typical running times and the variability of running (Strathman et al. 2003). For some properties, the additional times for a route, both of which contribute important infor- waiting time resulting from bus bunching has been standard- mation to the schedule writing process (Levinson 1991). ized into performance measures that are reported to senior Time-location data recovered by AVL systems can also be management (see the CTA and TriMet case studies, Appen- used to assess deviations in the actual delivery of service relative dices A and C, respectively). to the schedule. Analyses of such deviations across time points AVL systems also record the location status of vehicles at at the route and system level are the basis for reporting on-time specified time intervals (ranging from 30 to 90 s). These are re- performance. The general practice in the transit industry has ferred to as poll data. Poll data are transmitted over the vehi- been to define service as being "on time" when departures from cle's radio system to the dispatching center. In addition to time points are no more than 5 min. late or 1 min. early (Bates supporting dispatch functions, AVL poll data are used in real 1986). time to support software that broadcasts vehicle arrival time The pattern of depart times recorded by AVL for successive estimates through an agency's website or automated phone vehicles can also be compared against the scheduled head- system. While real time use of AVL poll data has benefited dis- ways to assess the degree of regularity in the service that is de- patchers and customers, archived poll data are of limited use- livered to customers. Especially in frequent service situations, fulness to market researchers because they cannot be easily

OCR for page 11
14 joined with other ITS data (given the absence of common lo- amount. Magnetic stripe card IDs can be registered, making cational references). the transactions of defined groups identifiable. For employer By counting boardings and alightings, and calculating pas- and university/school pass programs, this feature allows reli- senger loads, APC systems provide important data to market able documentation of usage and supports program develop- researchers and service planners. These passenger data, com- ment efforts (see the City of Madison Metro Transit case study piled at the stop, route, and system levels, establish the sampling in Appendix B). Smart cards have stored and recharge value ca- frames for rider surveys and provide the expansion factors pability and can be registered to individual users in connection needed to infer survey results to the riding population. The in- with personal accounts. Thus the transaction data from smart tegration of APC with AVL systems has greatly improved the card systems can identify individuals when they are registered. quality of passenger count data. Prior to AVL, passenger count Through integration with AVL systems, the transactions data from APCs were related to stops by inferring location from data for smart and magnetic card systems typically identify time and odometer stamps on the data records, a clumsy process riders' boarding locations. Analysis of the time sequence of that screened out much of the data recovered (Strathman and transactions, however, appears to be fairly successful in Hopper 1991). inferring riders' alighting locations (Rahbee and Czerwinski AVM technology recovers data on a vehicle's mechanical 2002). Similarly, when electronic fareboxes are integrated and electrical systems. While AVM data primarily support with AVL systems, summary totals for exit locations have maintenance functions, they are sometimes useful for market been successfully inferred from entry location totals (Navick research. AVM data on lift deployments, for example, and Furth 2002). document where riders with more serious mobility impair- Data elements for the remaining technologies covered in ments access the system, supporting surveys of this special this Guidebook are presented in Figure 2-4. Data from ticket population. AVM systems also record the transmission of sig- vending machines include each machine's location, the date nal priority requests when this function exists. These data and time of each transaction, the number and type of fare could be coordinated with corresponding signal system data items purchased, the price per item, the total value of each to assess time savings for riders. transaction, the method of payment; and the status of the MDTs and control heads allow operators to record prede- transaction. fined events by pressing an icon or button. Conceivably, the Web tracking software documents characteristics of range of "events" that affect customers' riding experience is Internet contacts, including the date, time, and duration extensive. These systems are capable of recording when and of each visit; the entry page, the page path through the site where such events occur. Event data commonly recovered by and exit page; and the identification of the referring website. these devices include overloads and pass-ups, safety and For websites that broadcast vehicle arrivals in real time, security incidents, fare evasions, medical emergencies, and tracking software record the route and stop location selected breakdowns (see the TriMet case study in Appendix C). and the time of the request. For websites with trip planning Electronic registering fareboxes record fare transactions software, tracking data cover the requests related to time of data, including fare type and transaction amount. Fareboxes travel, origin and destination locations, preferred travel that are equipped with keypads also allow operators to record options (e.g., shortest path or time, fewest transfers), and the fare media used and other predefined events (as described the final travel itinerary. Most websites include features that for MDTs and control heads). allow customers to communicate commendations, sugges- Among the advanced technology systems covered in Fig- tions, and complaints to the agency. In responding to these ure 2-3, electronic registering fareboxes have often been the communications, customer service staff usually documents first to be installed on a transit property's vehicles. When AVL characteristics of the communication (e.g., type, date, time, systems were subsequently acquired, their integration with route, location, and vehicle/operator identification, where existing fareboxes was sometimes not feasible or was not possible). pursued. Without AVL integration, farebox data are defined Tracking software for automated telephone systems doc- by vehicle and time stamps, resulting in similar location ument the characteristics of each call received, including the referencing problems experienced by APC systems in the pre- date, time, and duration of the call; the routing of the call AVL days. Where electronic farebox and AVL systems have through the system; and the origin of the call. Similar to the not been integrated, some effort has been made to identify Web, for phone systems with trip planning and real time transaction locations by "matching" farebox and AVL time vehicle arrival services, tracking software documents char- stamps (Cui 2006). acteristics of the information requested. Documentation Magnetic stripe and smart card systems recover fare trans- of customer communication of commendations, requests, actions data of the cardholders. For each transaction, magnetic and complaints via the phone system also occurs in a simi- card systems record the identity of the card and the transaction lar fashion.