Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Design of Advanced
Public Transportation
Systems
Seminar Report
Ashutosh Arun
QHS-Trainee
QHS-Trainee
AA
AA 2010 02
AA
2011
Design of Advanced Public Transport
System- Seminar Report
Ashutosh Arun
QHS Trainee
EXECUTIVE SUMMARY
This report is an attempt to understand the implementation of various APTS technologies in vogue in
the various cities of the world in the light of average Indian traffic conditions, with special emphasis
on the Bus Transit Operations.
Advanced Public Transportation Systems are a collection of information technology and software
applications along with essential physical systems viz. Geospatial Technologies (GIS, GPS etc.),
Automatic Vehicle Location, Advanced Communications, Transport Operations Software etc. These
technologies are applicable to almost all of the Public Transportation systems including Intermediate
Para Transit which currently ply on the urban roads in different parts of the world.
APTS have been widely adopted around the world and, with special importance to Bus Transit
Operations, been widely integrated into the Bus Rapid Transit (BRT) Systems of various cities, albeit
with varying rates of success. In India, an example of a BRT system utilizing some of these
technologies is the Janmarg in Ahmedabad, Gujarat.
This report focuses on the special issue of developing a general framework for simulation models for
designing APTS for Tier I and Tier II cities of India. Major inputs required for such models have been
discussed in detail.
Page | 2
OBJECTIVE
The objective of this report is to discuss in some detail the broad categories of inputs required for
developing simulation models which can be used for designing Transportation System for a Tier I or
Tier II Indian city integrating Advanced Public Transportation Systems.
Since it is a report developed for the seminar work as a part of the M. Tech. course of the author, it
would be highly appreciated of the reader to bear in mind the fact that it is a general overview of the
topics being considered and that the reader shall not expect to find very detailed technical
information. However, the list of references has been included for the benefit of the reader.
INTRODUCTION
The implementation of APTS technologies is transforming the way public transportation systems
operate, and changing the nature of the transportation services that can be offered by public
transportation systems. The goal is to provide public transportation decision-makers more information
to make effective decisions on systems and operations and to increase travellers’ convenience and
ridership. Also, it aims to equip the users with greater pertinent and updated information regarding the
system and hence empower them to make smarter decisions in their choices.
APTS technologies can be organized into five broad categories that describe the technologies'
relevance to transit applications. Each category is comprised of a variety of technology choices that
are available to help transit agencies and organizations meet travellers’ service needs while increasing
safety and efficiency.
Real-time management of bus and subway systems is now possible through the use of vehicle-based
fleet management systems. Advanced Public Transportation Systems (APTS) Fleet Management
systems present transit agencies with more effective tools for vehicle and fleet planning, scheduling,
operations, control of traffic signals, and monitoring of vehicle location. Following benefits can be
obtained through the use of these technologies:
According to Federal Transit Administration (FTA), there are six technologies falling under the
category of Fleet Management Systems:
CAD fixed route software falls into four primary categories, which are described in detail
below:
Page | 4
Expert systems for service restoration- Using real-time bus
operations information from AVL/CAD systems and the
existing operation rules and procedures, these expert
systems aim at providing optimum solution for restoration
of disrupted service.
Itinerary planning systems- Its function is to determine the
best way for a transit customer to travel from an origin to a
destination. Itinerary planning systems are indirectly
linked to CAD systems because they depend on the same
static route schedule data.
Service planning applications- Service planning
applications are used to analyze the efficiency of fixed-
route operations using data from AVL/CAD systems.
Service planning applications can perform the following
functions:
o Provide an average of running time between time Figure 1- Input Screen of an
points to determine which route segments are not Itinerary Planning Program
performing to schedule, by how much, and why;
o Calculate measures of headway and running time variability to determine where
service and scheduling should be adjusted for efficiency and as a measure of
customer service;
o Develop new performance measures on an as-needed basis; and
o Map the analysis for visual evaluation based on geographical referencing.
3. Communications Systems - They are used to transmit voice and data between vehicles and
operations centers, and to transmit commands between operators and technologies (e.g.,
signal pre-emption commands to traffic signal systems). The two-way voice radio system
used for fleet management and vehicle dispatching remains at the heart of most transit
operations. However, other communication technologies are becoming common; for example,
communications-based train control and short-range data links for traffic signal priority. For
bus transit operations, two types of communication technologies are used:
Mobile Voice and Data Communications Systems- It involves the broadcast of
information over Radio Frequency (RF) waves from a transmitter to receiver.
Short-Range and Other Communications- Communications systems, such as microwave,
Cellular Digital Packet Data, infrared, sonic, radio frequency, and fiber optic lines, are
excellent ways to transfer data and voice communications where long-range wireless
media are not suitable for one reason or another. E.g. Traffic Signal Priority Systems.
4. Geographic Information Systems- GIS provides a current, spatial, interactive visual
representation of transit operations. It is a type of computerized database management system
in which geographic databases are related to one another via a common set of location
coordinates. This allows users to make queries and selections of database records based on
geographic proximity and attributes. It is most often used for:
Transportation planning and modeling;
Demographic analysis;
Route planning, analysis, and restructuring;
Bus dispatch and scheduling;
Bus stop and facility inventory;
Ridership analysis;
Page | 5
Automatic vehicle location and tracking;
Fixed guideway facility management;
Paratransit scheduling and routing; and
Accident reporting and analysis.
State of California’s integration of APC (Automatic Passenger Counters), MDT (Mobile
Data Terminals) and AVL subsystems into a Web-based public information system is a good
example of this technology.
5. Automatic Passenger Counters- APCs typically use one of two counting technologies,
either treadle mats or infrared beams. Treadle mats, placed on the steps of the bus, register
passengers as they step on a mat and infrared beams (mounted either horizontally or
vertically), directed across the path of boarding and alighting passengers, register riders when
they break the beam.
6. Traffic Signal Priority Systems- Traffic signal priority is a strategy by which a particular set
of vehicles is given preference at traffic signals, either anytime they arrive at the intersection
or only under certain conditions (e.g., on-time status, amount of traffic at opposing
approaches).
To activate traffic signal priority for buses, a signal (via a sonic or optical pulse) is
transmitted from the bus to the traffic signal controller. Depending upon the phase the traffic
signal is in; the controller will either extend the current green phase or advance the timing of
the next green phase.
When used in combination, as illustrated in Figure 3, they form fleet management systems that help
transit agencies improve the efficiency and safety of transit service.
Page | 6
Figure 3- Complete Fleet Management System
APTS can provide travelers with more information in more different ways than ever before. Transit
information can be of a static nature such as route maps, schedules, and fares, or dynamic information
such as route delays and real-time arrival estimates. It can include continuously updated freeway and
arterial traffic flow conditions. All data is integrated into an automated, cohesive traveler information
database that can be accessed by travelers through Web sites and kiosks as well as e-mail alerts,
television/radio broadcasts, automated telephone systems, hand-held computing devices/mobile
phones, and variable message signs.
Agencies with AVL systems are able to provide real-time in-terminal or wayside transit
information. The primary devices for in-terminal and wayside systems are:
Video monitors for large amount of information and where flexibility in using graphics,
fonts, and color is needed. For example, a video monitor would be well-suited for a
display at the entrance to a station indicating which berth buses serving specific routes are
located.
Variable message signs for a relatively small
amount of information to be displayed at a
considerable distance from transit users, such as the
centralized real-time display. Also, they are
resistant to vandalism and can be used at bus stops. Figure 5- Public Information System
using VMS
These may be supplemented with audio announcements
of the displayed information. Real-time in-terminal and wayside information systems require
a communications link to a central computer system that provides the information about
upcoming arrivals.
Page | 8
The main components of systems using smart card based technology are:
A fare payment system - the infrastructure used to receive value from the fare payment
media and/or check the validity of the media for the current transit trip;
A fare distribution system - the infrastructure used for the distribution of the payment
media, as well as the distribution of the value that is loaded onto the fare media; and
Clearinghouse and back office processing systems - infrastructure used to capture and
process transaction data generated by the fare payment and distribution systems.
Transportation Demand Management (TDM) is a term applied to a broad range of strategies that are
intended to change traveler behavior for the purpose of reducing or reshaping use of the transportation
system. TDM strategies employ ITS technologies to:
Facilitate and increase the use of public transportation, including carpooling, vanpooling,
walking, bicycling, and telecommuting;
Compress work weeks;
Apply congestion pricing programs; and
Manage parking and apply demand pricing.
Three TDM strategies that utilize ITS technologies are discussed in this chapter:
IVI technologies reduce the probability of motor vehicle accidents through the use of vehicle controls
and driver warnings. The transit IVI platform provides an opportunity for evaluation of technologies
that improve the efficiency of the fleet and provide more safety:
Page | 10
CHAPTER 2: MODEL REQUIREMENTS FOR APTS SIMULATION
In order to simulate APTS applications in bus transit, it is necessary to represent bus transit operations
at a level of detail that supports the operational characteristics of the technology or system of interest.
For example, AVL/CAD systems that monitor bus performance and determine holding and
dispatching solutions to schedule deviations cannot be simulated in a model that does not represent
the bus transit schedule. The purpose of this chapter, then, is to summarize the requirements for a
microscopic traffic and transit simulation model to be able to simulate APTS.
At the core of a good microscopic traffic simulator are sophisticated driver and traveler behavioral
models that capture the complex interactions between vehicles and between vehicles and traffic
control and information systems. Similarly, APTS simulation should be based on a detailed, veritable
representation of bus transit operations.
Following is a diagram of the various APTS strategies discussed in Chapter 1 and the implications
they have with respect to bus transit operations and operations simulation.
Page | 11
Figure 9- APTS impacts on transit operations and
implications for simulation
In the diagram, the interactions between the model requirements can be seen, where the APTS enable
a variety of real-time operational strategies (e.g. holding and dispatching) that directly affect transit
vehicle movements, provide valuable input to planning applications that lead to better transit system
design (e.g. improved scheduling and route planning), and allow the sharing of real-time performance
information with travelers to influence demand and improve passenger level of service (e.g. route
choice).
The transport system variables that are more or less constant, such as routes, schedules and passenger
demand, are products of operations planning or passenger trip planning applications and best
simulated by way of input to the model. While, APTS that interact with real-time operations and that
provide real-time information to passengers can be represented within the simulation using models
that capture interactions
Page | 12
TRANSPORT SYSTEM REPRESENTATION
The transport system representation as an APTS modeling requirement can be divided into three parts:
Transport network
Schedule design
Fleet Assignment
Transport Network
The representation of the transport network includes the links, or paths, in the network that make up
bus routes, the designs and locations of bus stops along those routes, and the designs and locations of
other bus transit facilities, such as bus lanes, in the network.
This representation is critical to the interaction between the transit service and the passengers and the
surrounding traffic environment. For example, mixed traffic, as opposed to bus lane, transit routes
involve complex interactions between different modes. Bus stop design elements (e.g. single vs.
multiple berth stops) can have considerable impacts on vehicle operations at stops. Also, bus stops
that are located in the general traffic lane, as opposed to those that are removed in a wayside bay, will
require different bus operator maneuvers and stimulate different behaviors from other drivers in the
network.
System-wide
Route segment level
Bus stop level
Bus operations at the bus stop-level are uniquely separate from, but not independent of, route
segment-level operations. Passenger waiting times, bus dwell times, and boarding, alighting and
crowding phenomena occur at the bus stop level. The composite effects of dwell times at a series of
stops, traffic congestion and intersection delays, in turn, may be observed at the route segment level.
Schedule Design
The schedule design determines how and when buses serve the transport network. It depends upon
various types of bus services, such as bus rapid transit, fixed route services and demand responsive
services. Pine (1998) identifies three components of the transit schedule:
Route structure
Service frequencies
Service timing
Route structure refers to where the bus travels in the performance of its assigned trips and relates to
the transit network representation. However, service frequency refers to how often a bus passes a
given stop on the route, and service timing refers to when a bus arrives at a particular location on the
route.
Fleet Assignment
A vehicle assignment is defined as the work assignment given to a single transit vehicle for the
duration of a service workday. In the context of simulation, however, it may be considered the total
work assignment given to a transit vehicle for the course of the simulation.
Page | 13
For networks where the whole of a bus route is modeled, the work assignment might involve multiple
roundtrips on the route. Furthermore, the work assignments including trips where a single vehicle
serves more than one route, as is done in practice, can also be modeled.
When and how frequently buses arrive at specific locations along the route is a function of the
schedule and also susceptible to various random disturbances. These random variables include the
prevailing traffic conditions on and adjacent to the route (i.e. congestion), traffic control and
passenger demand (i.e. boarding and alighting passengers determine dwell time).
This interaction between the vehicle (bus in our case) and its environment can be taken up in two
parts: behavior between stops and behavior at and near stops.
Driving behavioral modeling is dominated by acceleration and lane-changing models, which are
typically complemented by more detailed models of gap acceptance, merging, and yielding. The
acceleration models for bus operators and other drivers might be assumed to be identical. Lane-
changing behavior, on the other hand, is fundamentally different between bus operators and other
drivers, and simulation models should reflect this difference.
Lane-changing theory assumes that each driver aims to minimize his or her travel time, and makes
discretionary lane changes based on the perceived utility of the alternative lanes, which is a function
of the relative speed of the vehicles in the target lane(s).
A bus operator might make discretionary lane changes between stops to increase travel speed
according to the same decision-making processes as other drivers, albeit with a preference toward the
lane with bus stops. However, mandatory lane-changing decisions, like changing into the extreme left
lane near the bus stops, will largely govern the bus operator lane-changing along a route. Thus the
location and spacing of stops along the route dictate to a great extent bus operator lane-changing.
Other factors affecting bus operator lane-changing that do not affect other drivers, such as the
presence of bus lanes and HOV lanes. Likewise, there are factors that affect other drivers but not the
bus operator. For instance, the private auto drivers traveling behind a bus in a lane that contains bus
stops will attempt to move out of the lane and overtake the bus in anticipation of the bus’ routine
stopping and starting at those stops.
Some simulation packages model the variation in familiarity with the network among the driving
population which governs various behaviors such as how far in advance of a turn or exit from the
current roadway a driver changes lane in order to make that turn or exit. A failure to capture bus
operator familiarity might cause a simulation model to overstate congestion when buses make late
lane change maneuvers.
Page | 14
Dwell Time is the period of time during which a bus is stopped at a bus stop to serve passengers and
consists of dead time and service time. Dead time is the sum of time spent stopped with the doors
closed and the time spent to open and close the doors. The service time is the span of time during
which the doors are open for passengers to board and alight. Dwell time depends upon the weather,
bus stop design and passenger demand.
Vehicle movement near bus stops can also have a serious impact on the surrounding traffic stream and
on vehicle progression. This particularly includes pulling out of bus stops. Unrealistic, excessive
delays may arise if a simulation model does not reflect the way drivers in the adjacent traffic stream
yield to exiting buses, and thus may cause undue disruption of the bus’ progression. The yielding
behavior of other drivers, in turn, can have significant consequences with respect to congestion in the
general traffic stream.
Passenger demand is generally considered to be random. For instance, passenger demand can be
highly variable at the route level, the sub-route (or route segment) level and the bus stop level. The
geographic distribution of passenger demand is subject to the local land use patterns and the locations
of activity centers along a route. Passenger demand may have considerable temporal variability. For
example, passenger demand might vary by time of day (e.g. peak and off-peak) and day of the week
(e.g. weekday vs. weekend) and is subject to spiking due to special events (e.g. World Cup matches).
Large passenger demand (boarding and alighting passengers) at a bus stop and crowding on the bus
might cause delays at the stop, thus preventing the bus from departing on schedule. A lack of demand
at a bus stop, however, in the absence of dispatcher intervention or operating procedures that call for
holding, might cause the bus to depart early and therefore get ahead of its schedule.
The number of passengers boarding and alighting depends on passenger arrival patterns, which is
usually a function of the type of service. For example, it is generally assumed that transit passengers
tend to arrive more randomly as the service becomes more frequent and irregular. On the other hand,
as the service becomes more regular and infrequent, passengers tend to rely more heavily on the
schedule and thus time their arrivals at stops closer to the scheduled vehicle arrival time in order to
minimize waiting time.
Jolliffe and Hutchinson (1975) divided transit passengers into three categories:
The proportion q who arrive coincidentally with the bus and thus have no waiting time.
The proportion p who are familiar with the schedule and arrive close to the vehicle arrival
time and wait on average wmin.
The proportion (1-q)(1-p) who arrive randomly and wait on average wrand.
When passengers arrive randomly, wrand = µ(1+ µ2/σ2)/2, where µ and σ are the mean and standard
deviation of the time headway between buses, respectively.
Based on measurements taken at bus stops in London, Joliffe and Hutchinson estimate p as a function
of the service characteristics: p = 1- exp (-λg)
The value of g is the potential to reduce waiting time, and increases as the time headway between
buses increases. Therefore, p increases when bus services become more infrequent.
Page | 15
The value of λ was determined to be 0.131 and 0.015 for peak and off-peak conditions, respectively,
confirming a priori expectations that, during the peak, more passengers are familiar with the schedule
and arrive so as to minimize waiting time.
This dynamic interaction between the passenger and the service is the subject of a host of APTS
applications, including traveler information systems, and has significant operational implications with
respect to APTS that mean to monitor and control bus operations in real time.
APTS REPRESENTATION
Various APTS technologies, whether for collecting information, applying information (e.g. for real-
time control) or disseminating information, differ widely in their designs and their operations. Some
APTS are designed to function offline and are not directly involved with transit operations in real-
time. Among these offline technologies are Itinerary Planning Systems (fleet management, transit
operations software) and some automated service coordination applications. The outcomes of these
offline applications can be used as input to a simulation model.
The online APTS technologies that make up the bus surveillance system serve as the link between the
supply and demand components of the bus transit system. Surveillance includes the sensor
technologies, and their governing logic, used by transport service providers to monitor the
performance of the system. These sensor technologies might include installed roadside bus sensors or
automated vehicle location technologies (e.g. GPS) or communications technologies that allow the
bus to transmit information to the TMC.
Using the surveillance system, transport service providers can monitor bus progression and make
informed decisions regarding real-time control of each bus’ movement and behavior (e.g. dispatch or
hold at a bus stop). The surveillance system may also be used to generate real-time input to various
APTS control strategies, such as conditional bus signal priority, and various information systems,
such as in-terminal/wayside traveler information.
In order to simulate APTS at the operational level, a traffic model must be able to mimic the
functionality of the technologies (e.g. where information is generated and how it is shared) as they
operate in the real world rather than the technology’s own intrinsic capabilities.
A generic and flexible bus operations model should be able to replicate the types of information (e.g.
location, speed, load) that the technologies produce and mimic the mechanism for sharing that
information between the components of the bus system (e.g. TMC, control devices, vehicles).
MEASURES OF EFFECTIVENESS
When simulating APTS, it is important to consider the benefits and costs of implementing a particular
APTS application. This might involve cost-benefit analysis at various levels of the transport system
including:
System level
Route segment level
Bus stop level
Vehicle level
Passenger level
Page | 16
The performance measures may be qualitative or quantitative. Qualitative factors include, for
example, passenger comfort, safety, and amenities at bus stops. Quantitative factors might include
monetary considerations, such as cost savings and revenue increases from increased ridership, or
service delivery measures, such as on-time performance and headway adherence.
The bulk of expected APTS benefits are quantitative gains that accrue either to the service provider
(e.g. cost savings, revenue increases) or to the passenger (e.g. reduced waiting and in-vehicle time).
APTS also produce benefits that occur in different parts of the network, such as at bus stops (e.g.
dwell time reduction), along a route segment (e.g. travel time, headway variability) and at the system
level (e.g. transit vs. auto travel times).
Therefore, performance measure output from traffic simulation should include data about the different
elements of the system in order to draw meaningful conclusions about the performance of APTS.
An example of measures of effectiveness for evaluating Traffic Signal Priority (Dale et. Al., 1999) is:
Bus Travel Times The time it takes a bus to travel the length of a
route or route segment
Bus Schedule Reliability The use of travel time variability (standard
deviation) as an indicator of reliability
Intersection Bus Delay Average delay to buses at an intersection
Average Person Delay Intersection control delay, intersection bus delay,
average automobile occupancies and bus loads to
determine delay in seconds/person
Vehicle Emissions CO and NOx emissions on a segment basis
Accidents Transit vehicle accident frequency as a safety
measure
Page | 17
CHAPTER 3: APTS IMPLEMENTATIONS
APTS have been and are being implemented in many developed/ developing countries in the world
viz. the US, Canada, Japan, Brazil, China etc. A FTA report Advanced Public Transportation System
Deployment in the United States- 1999, reveals that 212 transit agencies operate 361 service types
employing some form of advanced communications. Similar figures are available for other
technologies such as;
In India, a glowing example of the implementation of APTS is the BRT system in Indore city. The
following figure shows the technologies in vogue there.
APTS deployment in India is catching up because of BRTS coming up in cities like Pune, Ahmedabad
and Mysore and also because of the increased awareness about the benefits of such systems among
the Transportation agencies of other states and cities such as Hyderabad, Delhi, and Bangalore etc.
The following table shows the APTS services in public buses being used in some of the afore-
mentioned cities;
Page | 18
References
Daniel J. Morgan (2000). A Microscopic Simulation Laboratory for Advanced Public
Transportation System Evaluation. M.S. Thesis, Massachusetts Institute of Technology.
Federal Transit Administration. Advanced Public Transportation Systems: The State of the
Art Update 2000.
Robert F. Casey and John Collura (1994). Advanced Public Transportation Systems:
Evaluation Guidelines. Final Report, Office of Technical Assistance, Federal Transit
Administration.
Dario Hidalgo and Aileen Carrigan (2010). Modernizing Public Transportation. EMBARQ,
the WRI Center for Sustainable Transport.
Detailed Project Report- ITS Solutions, Transit Signal Priority and Automated Fare
collection Systems, Indore BRTS.
Page | 19