Intelligent Urban Traffic Control System

KKKA6423

Assignment -2-

Supervisor

Prof. Dr. Riza Atiq Abdullah OK

Prepared by: Sarah hazim P 65407

rasha salah ahmed P64799

6. April .2013

Introduction

Urban traffic control (UTC) systems are a specialist form of traffic

management which integrate and co-ordinate traffic signal control over a

wide area in order to control traffic flows on the road network. Integration

and co-ordination between adjacent traffic signals involves designing a plan

based on the occurrence and duration of individual signal aspects and the

time offsets between them and introducing a system to link the signals

together electronically. A traffic responsive signal control system is a means

of adjusting the traffic signal settings (cycles, green splits and offsets),

which optimize a given objective function, such as minimizing travel time or

stops, in real-time based upon estimates of traffic conditions. There are

many different UTC systems in operation around the world, but they can

provide the basis for an extended control system, generally termed Urban

Traffic Management and Control (UTMC).

UTC systems can be used to obtain better traffic performance from a road

network by reducing delays to vehicles and the number of times they have to

stop. UTC systems also can be used to balance capacity in a network, to

attract or deter traffic from particular routes or areas, to give priority to

specific categories of vehicles such as public transport or to arrange for

queuing to take place in suitable parts of the network.

Demand impacts usually reduce travel time, but reduced travel times and

good network performance may increase road capacity. This may cause a

shift in demand towards car use. UTC systems may not make a positive

contribution to all policy objectives.

Split, Cycle and Offset Optimization Technique (SCOOT):

The Transport Research Laboratory (TRL) in collaboration with UK traffic

systems suppliers developed the SCOOT (Split Cycle Offset Optimization

Technique) urban traffic control system. SCOOT is now co-owned by Peek

Traffic Ltd, TRL Ltd and Siemens Traffic Controls Ltd. Early systems were

tested in the late 1970’s in Glasgow. The development of SCOOT for

general use was carried out in Coventry with the first commercial system

being installed in Maidstone in 1980. SCOOT is now used in over 170 towns

and cities in the UK and overseas.

SCOOT is a fully adaptive traffic control system which uses data from

vehicle detectors and optimizes traffic signal settings to reduce vehicle

delays and stops. There are a number of basic philosophies which lead to the

development of SCOOT. One of these was to provide a fast response to

changes in traffic conditions to enable SCOOT to respond to variations in

traffic demand on a cycle-by-cycle basis. SCOOT responds rapidly to

changes in traffic, but not so rapidly that it is unstable; it avoids large

fluctuations in control behavior as a result of temporary changes in traffic

patterns.

SCOOT not only reduces delay and congestion but also contains other traffic

management facilities. For example, in 1995 a new facility was introduced

to integrate active priority to buses [link to bus priority] with the common

SCOOT UTC system. The system is designed to allow buses to be detected

either by selective vehicle detectors or by an automatic vehicle location

(AVL) system.

Impacts on demand

SCOOT is in use worldwide and has been shown to give significant benefits

over the best fixed time operation. The effectiveness of the SCOOT strategy

has been assessed by major trials in five cities (Wood, 1993; SCOOT

website). The results from the trials are summarized in the table below.

Comparisons of the benefits of SCOOT, against good fixed time plans,

showed reductions in delays to vehicles of average 27% at Foleshill Road in

Coventry - a radial network in Coventry with long link lengths. In Worcester

the use of SCOOT rather than fixed time UTC showed considerable saving

which was estimated to be 83,000 vehicle hours or 357,000 per annum at

1985 prices. The replacement of isolated signal control in Worcester by

SCOOT was also estimated to save 180,000 vehicle hours per annum or

750,000 per annum. In Southampton, economic benefit, excluding accident

and fire damage savings, amounted to approximately 140,000 per annum at

1984 prices for the Ports wood/St. Denys area alone.

Research by Bell (1986) suggests that SCOOT is likely to achieve an extra

3% reduction in delay for every year that a fixed-time plan "ages". Further,

the effects of incidents have been excluded from many of the survey results

to ensure statistical validity. Since SCOOT is designed to adapt

automatically to compensate for ageing and incident effects, it is reasonable

to expect that, in many practical situations, SCOOT will achieve savings in

delay of 20% or more.

In 1993 a SCOOT demonstration project in Toronto showed an average

reduction in journey time of 8% and vehicle delays of 17% over the existing

fixed time plans. During weekday evenings and Saturdays, vehicle delays

were reduced by 21% and 34%. In unusual conditions following a baseball

game, delays were reduced by 61%, demonstrating SCOOT's ability to react

to unusual events. (Siemens Automotive, 1995)

In Sao Paulo in 1997 a survey showed that SCOOT reduced vehicle delays

by an average of 20% in one area tested and 38% in another over the

existing TRANSYT fixed time plans. It was estimated that financial benefits

to Sao Paulo as a result of these delay reductions would amount to

approximately $1.5 US million per year. (Mazzamatti et al, 1998)

Impacts on supply

Field trials of bus priority using SCOOT survey were carried out in areas of

Camden Town and Edgeware Road in London in 1996. The Camden

network consisted of 11 nodes and 28 links. The Edgeware Road site was a

linear network consisting of 8 nodes and 2 pelican crossings. The bus routes

were surveyed for the periods 7:00 - 12:00 and 14:00 - 19:00. The results

show that greater benefits can be obtained where there is lower saturation

level. (Bretherton et al.1996).

Figure (1): The flow of information in SCOOT based UTC system.

UTOPIA (Urban Traffic Optimization by Integrated Automation) / SPOT

(System for Priority and Optimization of Traffic):

Is designed and developed by FIAT Research Centre, ITAL TEL and

MIZAR Automation in Turin, Italy. The objective of the system is to

improve both private and public transport efficiency. The system has been

fully operational since 1985 on a network of about forty signalized junctions

in the central area of Turin. The area also contains a tram line and control of

the trams is integrated within UTOPIA/SPOT (Wood, 1993).

UTOPIA/SPOT is now used in several cities in Italy and also in the

Netherlands, USA, Norway, Finland and Denmark.

The system uses a hierarchical-decentralized control strategy, involving

intelligent local controllers to communicate with other signal controllers as

well as with a central computer. Central to the philosophy of the

UTOPIA/SPOT system is the provision of priority to selected public

transport vehicles at signalized junctions and improvements in mobility for

private vehicles, subject to any delays necessary to accommodate priority

vehicles (Wood, 1993). The French PRODYN system and the German

MOTION system have some similarities to SPOT, but have not been used

outside their counties (Kronborg and Davidsson, 2000).

Impacts on demand

The improvements attributed to UTOPIA in Turin have been calculated a

previous traffic responsive control strategy rather than against a fixed time

system. Benefits of implementing UTOPIA were shown to give an increase

in private traffic speed of 9.5% in 1985 and 15.9% in 1986, following

system tuning. In peak times the speed increases were 35%. Public transport

vehicles, which were given absolute priority, showed a speed increase of

19.9% in 1985 (Wood, 1993).

SPOT was introduced in Scandinavia in the early 1990 (Kronborg and

Davidsson, 2000). In Oslo, Norway, SPOT started to be operated in four

intersections with high priority to public transport in 1996. Only traffic

parallel with the tram routes was evaluated and had good results (15%

reduction in travel time).

Impacts on supply

UTOPIA/SPOT has been explicitly designed with public transport vehicle

priority in mind (Wood, 1993). Buses and LRT vehicles are given absolutely

priority at junctions, subject to the accuracy in forecasting their arrival time.

In Turin LRT are given higher priority than buses because they have more

passengers but extra priority can be assigned on a vehicle by vehicle basis if

required.

Total benefits of UTOPIA-SPOT:

UTOPIA-SPOT offers the network manager the following benefits:

- Keeps the flow going

- Manages timely public transport

- Fully adaptive, adjusts to the traffic situation

- Realizes strategic traffic policy objectives.

- Dynamic priority levels for public transport vehicles.

- Tuned and tested in lab situation before installation on-site.

- Open communication infrastructure.

Gaps and weaknesses

Many papers or reports on UTC systems evaluated only the impact on

efficiency such as reduction in journey time, delay and stops compared with

previous types of system. However, reducing travel times can increase road

capacity, and increasing capacity over a significant area may cause a shift in

demand towards car use and increase car traffic volume. The potential for

the benefits of UTC systems to be eroded by induced traffic needs to be

borne in mind. Relatively little information is available on environmental or

safety benefits.

Suffolk County Accessible Transportation (SCATs):

Suffolk County Transit is the provider of bus services in Suffolk County,

New York on Long Island in the United States and is an agency of the

Suffolk County government. It was founded in 1980 as a county-run

oversight and funding agency for a group of private contract operators which

had previously provided such services on their own. While the physical

maintenance and operation of the buses are provided by these providers,

other matters ranging from bus purchases to route and schedule planning to

fare rules are set by Suffolk Transit itself.

Though serving the entirety of Suffolk County, the one exception is in

Huntington, located in the northwestern part of the county, where that town's

private operator declined to join Suffolk Transit. Instead, Huntington took

over that town's system which became Huntington Area Rapid Transit, or

HART. Most of HART's routes do connect to both Suffolk Transit and

Nassau Inter-County Express and one can transfer between HART and

Suffolk Transit fairly easily. In addition, the village of Patchogue has its

own local bus service

Suffolk County Accessible Transportation (SCAT) is Suffolk Transit's

federally-mandated paratransit service for ADA-eligible passengers with

disabilities. SCAT service is available Monday through Friday, 6:00 AM to

8:30 PM and Saturday, 7:00 AM to 8:30 PM. The fare is $3.00.

Fare:

The current Suffolk County Transit base fare for most one-way local bus

travel is $2.00. For seniors and the disabled, the base fare is $0.75;

personal care attendants (PCA) may ride for free when traveling with

seniors or the disabled. Students with school-issued identification pay a

reduced fare of $1.25. Children under five years of age are free, with a limit

of three children for every adult. On routes S92 and 10C, the base fare is

$2.25[4]

Fare payment is conducted with the use of coins or paper currency, and

must be exact. Bus transfers cost an additional $0.25, and must be requested

and paid for upon boarding the bus. These transfers are valid for two hours

after issue and can be used on Suffolk County Transit connecting routes, or

to Nassau Inter-County Express (NICE) connecting routes with a special

transfer request slip (transfers to NICE require payment of a "step-up" fare.

Intelligent Traffic Adaptive Control Area (ITACA)

The ITACA adaptive traffic control system its application to traffic control

in the Spanish cities of Madrid and Barcelona. ITACA offers real-time

response to current and future traffic flow demands, and brings 'intelligence'

to fixed-time pattern control approaches. It incorporates: (1) an adaptive

system, which is used to evolve the best plan at each junction; and (2) an

expert system, which can use all the adaptive system's data and predictions

to obtain a global solution for the total traffic plan. This solution is

communicated to the adaptive system by a sophisticated use of importance

(weight) factors. The adaptive system has cycle, split, and offset optimizers,

and uses profiles to update the road network model. The model's

components include: (1) queue lengths; (2) congestion indicator; (3) load;

and (4) saturation flow modifier. The expert system is an optional part of

ITACA, which uses the model's current network data and its rules to adjust

the weights of each traffic movement. Its most obvious use is to avoid

secondary congestion, the blocking of junction exits by downstream queues.

It is expected that congestion strategies will develop differently for each

network, and depend strongly on users. Any number of overlapping

concurrent strategies can be implemented. For the covering abstract see

IRRD 877920.

ITACA recommendation:

From the exchange of knowledge with the project partners and the

discussions at ITACA meetings in Brabant some recommendations were

formulated to incorporate in a road map to Sustainable Mobility in Noord

Brabant:

Transition from value-added chain to a value-added network;

Stimulate and promote E-biking for shorter distances and develop

and promote sharing vehicles.

Make sustainable road transport an integrated part of city investment

plans

A market approach for mobility (supply and demand) combined with

government’s responsibility for market regulation

Experience, convenience, comfort, and personal safety should

become the first principles for development en innovation.

Max band: Max band is a bandwidth optimization program that calculates

signal timing plans on arterials and triangular networks. MAXBAND

produce cycle lengths offset speeds and phased sequences to maximize a

weighted sum of bandwidths. The primary advantage of MAXBAND is the

freedom to provide a range for the cycle time and speed. The lack of

incorporated bus flows and limited field tests are disadvantages of

MAXBAND.

Now day’s microcomputers are as commonly available as the electronic

calculators of the 70s and, while more expensive than calculators, are easily

within the economic reach virtually; to every transportation professional in

most locations throughout the world.

Developers of computerized traffic tools, such as the U.S. Department of

Transportation and some state Departments of Highways and Transportation,

universities and private organizations have promulgated a substantial suite of

software tools for every phase of transportation planning and engineering in

the past decades. The Federal Highway Administration (FHWA) and Federal

Transit Administration (FTA) have even set up microcomputer software

distribution and support centers to help get the products to users.

Currently for example, the Center for Microcomputers in Transportation

(McTrans), lists over 475 software tools in these functional areas:

Construction management;

Highway design, pavements, bridge design and hydraulics;

Maintenance;

Safety and accident records;

Surveying;

Traffic engineering;

Transit; and

Urban transportation planning.

RONDOn (Rolling horizon based Dynamic Optimization of signal

control):

Is a newly developed real-time traffic adaptive signal control system that

aims to reduce the response delay against the sudden changes of traffic flow.

RONDO project started in 1998. Since then continuous enhancements to

RONDO have been undertaken. Now, RONDO is challenging the new

problems, which are to promote traffic safety and to protect the

environments with keeping traffic efficiency. In this paper, the latest

additional functions are introduced to solve these problems. A plan to install

the pilot system at the beginning of 2001 is described.

Application

Rondo uses a feedback loop to govern the behavior of traffic in the network

core. It manages the flows that originate and terminate between various PoPs

(Points of Presence) in the network by directing these flows into the multiple

pathways that are created using MPLS Label Switched Paths. These LSPs

serve as conduits through the network that are unaffected by the local

optimization strategy of shortest path routing. Rather, Rondo optimizes

performance based on global traffic considerations in the network.

System Components

Rondo is composed of the major parts shown in Figure 2 above.

In the remainder of this paper, we will describe each element with emphasis

on the data collection subsystem and the analysis engine.

Physical Network

The experimental network is a set of 10 MPLS-enabled counters and Inter

connections patterned after a much-scaleddown representation of a major

service provider’s network backbone as depicted on their web site. We note

that the provider has 2500 PoPs worldwide so our model has only rough

equivalence to reality. However, even with only ten routers, our network

exhibits complex and often fascinating behaviors. Routers are connected

with 10-megabit links, which makes possible the creation of realistic

overload conditions. Each router models a PoP (Point of Presence) on the

network where customer nodes are attached. In Rondo, each node attached to

a PoP is a PC that sends and receives packets.

The network uses a combination of Cisco® 3620 and 3640 series routers.

The release of Cisco’s IOS (Internet Operating System) available on our

routers allows only destination - based selection of MPLS tunnels. -Cisco is

a registered trademark of Cisco Systems, Inc. Upgrades will ultimately allow

selection of the tunnels based on other parameters in the IP packet.

Programmable Load Generators and Loading Strategy

We use a collection of PCs programmed to generate time-varying loads

similar to those expected in an operational network. Background network

traffic on the network is constant in time and is generated by commercially

available packet generators. Loads are carefully crafted to cause a buildup of

congestion that does not have an overall steady state solution, and are

designed to stress the given physical topology.

Data-Collection System

The data-collection system uses a variety of devices and techniques to

monitor the conditions in the network. These include both active and passive

methodologies that capture such characteristics as throughput, loss, delay

and jitter. Data collection, a key part of Rondo, uses an extensible

architecture to provide rapid processing of data under time constraints for its

collection, reduction and transmission. Data flow from the network probes

through the collection system to the analysis engine with little latency and to

archival storage at a lower priority. Data are retained in a database system

for other applications such as service-level management that do not require

rapid data processing. We describe this part of the system in detail below.

Data Model and Database

Rondo uses the database for a variety of classes of information including

physical and logical network topology, configuration information and

archived measurement data. The algorithms, displays and other components

are driven by the information described by this model, and as such, the

organization of this model is crucial to the effectiveness of Rondo. The

model, which is important for other applications, is realized in a relational

database. The most important function of the database is to hold the state of

the network topology, which changes as the system reroutes LSPs to

alleviate congestion. The analysis and reroute engine periodically updates

the topology as the network is reconfigured.

Analysis and Rerouting Engine

This element of the system contains techniques for detecting congestion in a

network and altering the existing traffic flows to eliminate an overload

condition. The engine is designed to focus on more than link utilization,

which is the most basic metric of network performance. Utilization indicates

the level of activity between network elements and is often viewed as a

measure of network congestion. This view is too simple when one considers

the classes of traffic that flow over an IP network. High utilization of a link

is one form of congestion, but others might include excessive delay, jitter or

high packet loss, all of which could happen at relatively low levels of link

utilization. These are measures of congestion that seriously affect proposed

services in next-generation IP networks, including voice and video. The

engine is designed use any measurable quantity as an indication of a network

problem that needs correction.

MPLS Configuration and Control

Rondo relies on MPLS to form explicit paths through the core network.

Explicit paths allow precise control over the placement of traffic flows

within the routed domain of Rondo. All traffic in Rondo flows through

explicitly routed MPLS tunnels, which specify each node along a path from

the ingress to egress routers. The network configuration is initially optimal

in the sense that all tunnels travel via the shortest path in the network. Once

established, packets enter the MPLS tunnels as a function of their destination

address and are delivered to the egress router.

Rondo thus uses MPLS as a mechanism for packet forwarding that is not

directly aware of quality of service. Mixing packets with different levels of

quality of service in an LSP is possible though but limits the effectiveness of

available controls. Once the initial explicit paths are established, the analysis

and reroute engine operates to reroute packets through a path established by

a new MPLS tunnel, which may no longer be the shortest path. This action

currently takes place via IOS commands that are issued from the controller.

When MPLS traffic-engineering MIBs become available, the controller will

use SNMP to establish the new routes.

System Operation

The analysis and rerouting engine is in overall control of the system. The

engine communicates with the data collection system to establish a schedule

of network measurements. As the data collection system takes each

measurement, it notifies the analysis and rerouting engine of the presence of

new data. The engine combines the new data with the current system

configuration and previous data to decide on the appropriateness of rerouting

an MPLS tunnel. If a move is appropriate, the analysis engine reconfigures

the network through the LSP configuration control and updates the network

state in the database.

As we discuss in the following, the route of the new MPLS tunnel does not

necessarily preserve overall network optimality. Rather our goal is to reroute

traffic as quickly as possible to minimize the congestion at the expense of

achieving a theoretical optimum over the whole network. Global

optimization might imply moving many or even all the routes in the

network. The strategy in Rondo is to move from one to a few MPLS tunnels

over a period of a few minutes with minimal disruption to network traffic.

Conclusions

The particular techniques proposed are experimental and not yet

mainstream, especially when proposed for such a large, on-line, application.

The pro-active and re-active nature of agents can be a helpful paradigm in

intelligent traffic management and control. Further (real-life) tests on a

control strategy, based on intelligent and autonomous agents, are necessary

to provide appropriate evidence for operational use as relatively little is

known about the global behavior of these intelligent agent systems when

they are scaled up to deal with more realistic problems.

As this research is still ongoing we hope, in the end, to demonstrate that an

integrated dynamic urban traffic control system based on agent technology

can adapt and respond to real world traffic conditions in real-time. A

working prototype of such a system should give appropriate evidence on the

usability of AI agent based control systems.

Signal control systems that have the capability of optimizing and adjusting

the traffic light settings are able to improve the vehicular throughput and

minimize delay through appropriate response to changes in demand patterns.

With the introduction of two un-coupled loops, whether agent technology is

used or not, a different theory of traffic control can be met.

Artificial agents are a metaphor to be used for theoretical and

implementation purposes. Primarily results indicate that given an automated

control strategy implemented in the traffic signaling devices we can get a

system that makes better use of the capacity of the intersection. It has been

shown that control systems based on agent technology can adapt and

respond to changing conditions in real-time and in the meantime making

better use of the infrastructure.