ec1-dynamic traffic light control system

7/30/2019 EC1-dynamic traffic light control system

1/12

A dynamic and automatic traffic light control expert system forsolving the road congestion problem

W. Wen

Department of Information Management, LungHwa University of Science and Technology, Taiwan, ROC

Abstract

Traffic congestion is a severe problem in many modern cities around the world. To solve the problem, we have proposed a frameworkfor a dynamic and automatic traffic light control expert system combined with a simulation model, which is composed of six submodelscoded in Arena to help analyze the traffic problem. The model adopts interarrival time and interdeparture time to simulate the arrivaland leaving number of cars on roads. In the experiment, each submodel represents a road that has three intersections. The simulationresults physically prove the efficiency of the traffic system in an urban area, because the average waiting time of cars at every intersectionis sharply dropped when the red light duration is 65 s and the green light time duration is 125 s. Meanwhile, further analysis also shows ifwe keep the interarrival time of roads A, B, and C, and change that of roads D, E, and F from 1.7 to 3.4 s and the interdeparture times atthe three intersections on roads A, B, and C are equal to 0.6 s, the total performance of the simulation model is the best. Finally, accord-ing to the data collected from RFID readers and the best, second and third best traffic light durations generated from the simulationmodel, the automatic and dynamic traffic light control expert system can control how long traffic signals should be for trafficimprovement. 2007 Elsevier Ltd. All rights reserved.

Keywords: Simulation system; Traffic light control expert system; Radio Frequency identification; Traffic congestion

1. Introduction

Traffic congestion has been causing many critical prob-lems and challenges in most cities of modern countries.To a commuter or traveller, congestion means lost time,missed opportunities, and frustration. To an employer, con-gestion means lost worker productivity, trade opportuni-ties, delivery delays, and increased costs. To solve

congestion problems is feasible not only by physically con-structing new facilities and policies but also by buildinginformation technology transportation management sys-tems. A growing body of evidence proves that simplyexpanding a road infrastructure cannot solve traffic conges-tion problems. In fact, building new roads can actuallycompound congestion, in some cases, by inducing greater

demands for vehicle travel demands that quickly eat awaythe additional capacity. Therefore, many countries areworking to manage their existing transportation systemsto improve mobility, safety, and traffic flows in order toreduce the demand of vehicle use. By enhancing publictransport, route guidance systems, traffic signal improve-ments, and incident management, congestion can beimproved greatly. Of course, construction of new private

bus way, expressways, or subway to increase these growthfor easy travel has not kept pace. From a recent analyticalstatistics of the US department of transportation (2007), itis estimated that roughly half of the congestion is what isknown as recurring congestion caused by recurringdemands that exist virtually every day, where road useexceeds existing capacity. The other half is due to non-recurring congestion caused by temporary disruptions.Four main reasons of non-recurring congestion are: trafficincidents (ranging from disabled vehicles to major crashes),work zones, weather, and special events. Non-recurring

0957-4174/$ - see front matter 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.eswa.2007.03.007

E-mail address: [email protected]

www.elsevier.com/locate/eswa

Available online at www.sciencedirect.com

Expert Systems with Applications 34 (2008) 23702381

Expert Systemswith Applications
mailto:[email protected]:[email protected]


2/12

events dramatically reduce available capacity and reliabilityof the entire transportation system. Therefore, researchershave done many researches to increase capacity and removebottlenecks.

Schaefer, Upchurch, and Ashur (1998) developed a sim-ulation model for evaluating freeway lane control signing.

The simulation results show that lane control has littleinfluence on congestion. However, the region betweenheavy and medium traffic flow is sensitive to lane control.Chen and Yang (2000) and Chen and Yang (2003) havecreated an algorithm to find a minimum total time pathto simulate the operations of traffic light control in a city.Stoilova and Stoilov (1998) also built a simulation modelto measure the best of traffic lights to achieve low noise lev-els with optimal traffic management and environmentalpollution. Grau and Barcelo (1992) and Messmer andPapageorgiou (1994) discussed the minimum of queuelengths in different intersections. Meanwhile, to aid trafficmanagement systems, Nooralahiyan, Dougherty, Mcke-

own, and Kirby (1997) adopted a Time Delay Neural Net-work (TDNN) to classify individual traveling vehiclesbased on their speed-independent acoustic signature. Wenand Hsu (2005) designed a route navigation system witha new revised shortest-path routing algorithm and madea comparison of performance evaluation. Besides, manyresearches on how to avoid traffic congestion by usingthe shortest-path algorithm have been published Chabini(1998), Chabini (1997), Hoyer and Jumar (1994), Ikeda,Hsu, and Imai (1994), Ikeda and Imai1 (1994) andManiccam (2006).

Moreover, the widespread use of information technol-

ogy provides an opportunity to enhance the techniques ofexpert systems (ES), which help managers deal with fastchanging environment at a human expert with high-qualityperformance. Expert systems have a variety of applicationsin many areas. Additionally, researchers have attemptedto develop effective intelligent systems to assist managersin making decisions about how to solve various prob-lems Liao (2002), Liu (1997), Maniccam (2006), Sheu(2006), Xia and Shao (2005), Yang and Recker (2005).Wangermann and Stengel (1998) proposed an intelligentaircraft/airspace system that provides better system perfor-mance, redundancy, and safety by using the overlappingcapabilities of agents. Powerful and flexible multiple agentswith the function of principled negotiations are communi-cated each other. The system gives aircraft and airlinesgreater freedom to optimize their operations than theirhave now. Wen and Yang (2006) developed a dynamicand automatic traffic light control system for solving theroad congestion problem. They simulated a specific road,the Chung San North road in Taipei, Taiwan, to discusswhether a road simulation model can solve a congestionproblem. Findle, Surender, and Catrava (1997) developeda flexible and general on-line method to determine whetherthe phasing as an intersection in given traffic flow scenariosneeds a protected left-turn. In their study, a simulation

model and was constructed to reproduce the effect of per-

mitted left-turns at an intersection. For comparison pur-poses, a number of experiments were carried out. Theirnew approach has been proven to produce better intersec-tion performance than the 50,000 rule over a significantrange of traffic flows. Eriksson (1996) utilized a two-tierarchitecture, a clientserver model, to build a straightfor-

ward expert system which was coded in HTML embeddedJava Applet to communicate with a knowledge server inthe back end. Some user-interface operations such asmouse dragging, display, and field checking, were put intothe front-end machine that can provide rapid response. Fay(2000) described a railway dispatching system, which has aknowledge base in fuzzy rules of the IFTHEN type. Thesystem adopts fuzzy reasoning to obtain train traffic con-trol decisions. The study shows that by systematically mak-ing use of the knowledge of train dispatching, traffic qualitycan be improved and operation costs can be reduced. Theabove descriptions show that using an expert system com-bined with a traffic light control simulation model is a good

idea for solving congestion problem. Therefore, our studyfocuses on traffic signal improvements to improve trafficcongestion problem.

The remainder of this paper is organized as follows. Sec-tion 2 introduces the framework for the dynamic and auto-matic traffic light control expert system. Section 2.1describes the description of a simulation model for control-ling traffic signals. Section 2.2 gives definitions and nota-tions for the simulation model. In Section 3, simulationanalysis and results for improving road traffic problemsare illustrated. Finally, some important conclusions andfuture work are discussed in Section 4.

2. The framework for the dynamic and automatic traffic light

control expert system

The dynamic and automatic traffic light control expertsystem (DATLCES) is composed of seven elements: a radiofrequency identification (RFID) reader, an active RFIDtag, a personal digital assistance (PDA), a wireless net-work, a database, a knowledge base, and a backend server(see Fig. 1). The RFID reader detects a RF-ACTIVE codeat 1024 MHz from the active tag pasted on a car. Theactive tag includes a battery so that it can periodically

Fig. 1. A framework for dynamic and automatic traffic light control

expert systems.

W. Wen / Expert Systems with Applications 34 (2008) 23702381 2371


3/12

and actively transmit messages stored in the tag. Uponreceiving the data, the reader will save all information inthe PDA. After accumulating fix size of data, the PDA witha wireless card will connect to the backend server and storethem into the database in the server. The server uses thedata stored in the database to calculate maximum flow,

interarrival time, and average car speed. When all possiblecongestion roads and car speed are collected, these datawill be used as the input parameters of the traffic light con-trol simulation model built in the server. Upon getting thesimulation results, the DTLCES is able to automaticallyinfer and provide different alternatives in terms of varietiesof traffic situations and then set red or green light durationvia a traffic light control interface for improving the trafficcongestion problem. For easily performing knowledge rep-resentation and reasoning, all rules in the knowledge baseare presented as a set of if hantecedent clausesithen hconse-quent clausesirules. Basically, there are two types of reason-ing: backward chaining and forward chaining. Forward

chaining, which is a data-driven approach, starts from abasic idea and then tries to draw conclusions. It checksthe IF part of IFTHEN rules to find out whether the ante-cedent clauses of the rule is matched. As each rule is tested,the program can inference one or more conclusions. Con-versely, backward chaining is a goal-driven method.

It starts with a goal to be verified as either true or false.It then examines all the THEN parts of IFTHEN rules. Arule that contains this goal in its consequent clause will bechecked to know whether its antecedent clause is true. Iffails, the program searches for another rule whoseconclusion is the same as that of the first one. This process

continues until all possibilities have been tested. Ourapproach makes use of forward chaining.

To clarify the forward chaining procedure, severalexample rules in the knowledge base for controlling trafficsignals are shown below:

Rule: 1 if 3.0 < Interarrival_time then Singal_Type = 1Rule: 2 if 1.7 < Interarrival_time5 3.0 then Singal_

Type = 2Rule: 3 if 0 5 Ineterarrival_time 5 1.7 then Singal_

Type = 3Rule: 4 if Interarrival_time = Exception then Singal_

Type = 4Rule: 5 if Singal_Type = 1 then Red_light_duration =

65 and Green_light_duration = 95Rule: 6 if Singal_Type = 2 then Red_light_dura-

tion = 65 and Green_light_duration = 110Rule: 7 if Singal_Type = 3 then Red_light_duration =

65 and Green_light_duration = 125Rule: 8 if Singal_Type = 4 then Red_light_duration =

Manual and Green_light_duration = ManualRule: 9 if 0 5 Interarrival_time 5 1.7 then Interdeparture_

time = 0.6

The algorithm for controlling all the traffic lights is as

follows:

for(;;){if(3.0 < Interarrival_time)

{then Red_light_duration = 65 and Green_light_

duration = 95;

}endif;if(1.7 < Inte0rarrival_time 5 3.0)

{then Red_light_duration = 65 and Green_light_

duration = 110;}endif;

if(0 5 Interarrival_time 5 1.7){

then Red_light_duration = 65 and Green_light_duration = 125

and Interdeparture_time = 0.6;

}endif;

if(Interarrival_time = Exception){

then Red_light_duration = Manual and Green_light_duration

=Manual;}endif;exit(0);

}

Basically, the simulation model described in the nextsection will only be run once. Once we get three optimalalternatives, the best, second and third best traffic lightduration, and the interarrival time collected and computedfrom the server, the DTLCES uses the above algorithm toreason and control the traffic light signals. Notice that thetraffic light duration on roads A, B, and C is the same asthe green light duration on roads D, E, and F.

2.1. Description of the simulation model

The 6-road traffic control simulation model is composedof six submodels, A, B, C, D, E, and F, which simulate A,B, C, D, E, and F roads traffic conditions as shown Fig. 2.Submodels A, B, and C are similar and Submodels D, E,and F are similar. Therefore, we only explain the processesof submodels A and D (see Figs. 3 and 4). The rest of sub-models can be understood easily. In the model, we assumethat the interarrival time for each car is 1.7 s, the interde-parture time for a car passing the stop line at each intersec-tion is 1.2 s, and the time for passing a length of a car is0.41 s, which are physically observed in average. We alsosuppose that the traffic light signals at every intersectionon a road are set to the same duration and color. InFig. 3, there are four modules named Light control A1,

Light control A2, Light control A3, and Car arrival A.

2372 W. Wen / Expert Systems with Applications 34 (2008) 23702381


4/12

The process in the upper dash area of Fig. 3 controls thetraffic signal at the first intersection on road A. Light con-trol A1 generates an entity to control traffic light signal.Assign for light A1 clock timegets the current simulationtime. Prempt seizes the resource switch A1 that has first pri-

ority to get the resource. Delay for red light A1 sets theduration for red light. Release switch A1 releases theresource switch A1 for allowing cars to seize the resource.Finally, Delay for green light A1 sets the duration for greenlight. Like the process of Light control A1, the processes ofLight control A2 and Light control A3 are the same. So, it iseasy to derive from the same processes.

The lower dash area of Fig. 3 shows the moving processof cars in the first segment of road A. Car arrival A gen-erates each entity for representing a car based on the inter-arrival time. The block Queue presents the average numberor waiting time of cars in the queue, which cannot passthrough the first intersection on road A. The block Seizecatches the resource switch A1 for getting the right to pass-ing through the first intersection on road A. Seize space ofthe second segment A road takes the resource spaceA2 tocheck whether the second segment on the A road hasspace or not. Then, Assign for clock time passing intersec-tion A1 uses an equation, TNOW+ cross A1, to representthe time for passing the first intersection on road A. Delayfor passing intersection A1 is the time for a car passing thestop line at the first intersection on road A. Release switchA1 for next car frees the resource switch A1 for allowingthe next car to move. Assign for decreasing space A2reduces by 1 to calculate the space in the second segment

on road A. Delay for driving through the second segment A

road uses an equation, NA2*T, to represent how long itwill take to pass through the remaining space in the sec-ond segment on road A. Upon the traffic light at the firstintersection changing to red, a car entity holding switch A1needs to release the resource switch A1 (i.e., the Preempt

block in the upper dash area in Fig. 3 takes one unit ofa resource switch A1 away from the Seize block in thelower dash area that originally seizes it). The interruptedcar entity then will be sent to the Delay block (i.e., adummy block whose value is 0). The purpose of thedummy block is to continue the process for going toDecide clock light A1 Eqt clock cross A1. Next, Decideclock light A1 Eqt clock cross A1 examines if clock lightA1 is equal to cross A1, then delay the remaining time,which is 0, for passing the stop line at intersection A1,then go to Assign for decreasing space A2. If clock lightA1 is not equal tocross A1, then go to the module Decideremaining time for passing intersection A1 Eqt 0. If theremaining time for passing the stop line at intersectionA1 is equal to 0, then the car entity will be added intothe Queue block with the first priority by using the Insertblock. Otherwise, the car entity will go to the Delay blockon the right side of Decide remaining time for passing inter-section A1 Eqt 0 and delay the remaining time.

2.2. Definitions and notations of the simulation model

To illustrate how the system works, we have developed asix-road traffic simulation model, which contains nineintersections, by using Arena. Before giving example, let

us introduce the definition of notations as follows:

Submodel A

Submodel B

Submodel C

Submodel D Submodel E Submodel F

Traffic

Light A1

Traffic

Light D1

Traffic

Light E1Traffic

Light F1

Traffic

Light D2Traffic

Light E2

Traffic

Light F2

Traffic

Light D3

Traffic

Light E3

Traffic

Light F3

Traffic

Light A2

Traffic

Light A3

Traffic

Light B1

Traffic

Light B2

Traffic

Light B3

Traffic

Light C1

Traffic

Light C2

Traffic

Light C3

Fig. 2. A traffic control simulation.



5/12


6/12


7/12

clock cross i: Current simulation time of the car permittingto leave the stop line at the first intersection(TNOW) pluses cross I where i= A1, B1,C1, D1, E1, F1.

clock cross j: Current simulation time of the car arriving atthe stop line at the second intersection

(TNOW) pluses cross jwhere j= A2, B2, C2,D2, E2, F2.green i: Green light duration at the first intersection

where i= A1, B1, C1, D1, E1, F1.red i: Red light duration at the first intersection

where i= A1, B1, C1, D1, E1, F1.green j: Green light duration at the second intersec-

tion where j= A2, B2, C2, D2, E2, F2.red j: Red light duration at the second intersection

where j= A2, B2, C2, D2, E2, F2.green k: Green light duration at the third intersection

where k= A3, B3, C3, D3, E3, F3.red k: Red light duration at the third intersection

where k= A3, B3, C3, D3, E3, F3.

2.4. Resources

switch i: A resource for controlling traffic light signal orfor determining whether a car permits to passthe stop line at the first intersection wherei= A1, B1, C1, D1, E1, F1.

switch j: A resource for controlling traffic light signal orfor determining whether a car permits to passthe stop line at the second intersection wherej= A2, B2, C2, D2, E2, F2.

switch k: A resource for controlling traffic light signal orfor determining whether a car permits to passthe stop line at the third intersection wherek= A3, B3, C3, D3, E3, F3.

2.5. Queues

iQueue: A queue at the first intersection to store all wait-ing cars, which do not have rights to pass theintersection where i= A1, B1, C1, D1, E1, F1.

jQueue: A queue at the second intersection to store all

waiting cars, which do not have rights to passthe intersection where j= A2, B2, C2, D2, E2,F2.

kQueue: A queue at the third intersection to store all wait-ing cars, which do not have rights to pass theintersection where k= A3, B3, C3, D3, E3, F3.

3. Simulation analysis and results

Using Arena, we have built a traffic light control signalsimulation model, which consists of six submodels, A, B, C,D, E, and F. During the simulation analysis, initially a

warm-up time of 1800 s for reaching system stable and arun time of 10,000 s are set. Then, we set the green lightduration with 50 s and the red light duration with 50 s asthe first group. The rest of groups are selected based onboth red light and green light duration increased by 15 seach time (i.e., r = 50, g= 50; r = 65, g= 65; r = 80,g= 80; r = 95, g= 95; r = 110, g= 110; and r = 125,g= 125 s). So in Table 1, totally we inspect and analyze 6cases in terms of different light durations.

As we mentioned earlier, iQueue represents a queue atthe first intersection to store all waiting cars, which donot have rights to pass the intersection where i= A1, B1,

C1, D1, E1, F1. According to the result of this simulationanalysis in Table 1 the average waiting times in A1Queue,B1Queue, and C1Queue are very high (see Fig. 5). There-fore, most cars are waiting at the first intersection on roads

A1Q

ueue

A2Q

ueue

A3Q

ueue

B1Q

ueue

B2Q

ueue

B3Q

ueue

C1Q

ueue

C2Q

ueue

C3Q

ueue

0

500

1000

1500

2000

r125g125

r110g110

r95g95

r80g80

r65g65

r50g50

Fig. 5. The average waiting times for the same light durations (Roads A,B, and C).

Table 1The average waiting times for the same light durations (roads A, B, and C)

A1Queue A2Queue A3Queue B1Queue B2Queue B3Queue C1Queue C2Queue C3Queue

r50g50 1712.85 49.0432 37.8693 1712.85 49.0398 37.9105 1712.85 49.0413 37.9125r65g65 1685.33 48.4127 29.0645 1685.33 48.4121 29.0378 1685.33 48.4125 29.056r80g80 1723.46 49.2698 46.4928 1723.46 49.2663 46.5198 1723.46 49.2736 46.4816r95g95 1702.55 49.2105 49.2501 1702.55 49.2096 49.2302 1702.55 49.2021 49.2346r110g110 1740.85 49.3554 46.5865 1740.85 49.3503 46.6262 1740.85 48.3535 46.5965

r125g125 1743.5 49.0452 49.0754 1743.5 49.0138 49.0889 1743.5 49.0352 49.0729



8/12

A, B, and C. For the second and third intersections, thereare a few cars in the waiting queues. Therefore, based onthe average waiting times of A1Queue, B1Queue, andC1Queue, we can find that it is a better case when the greenlight duration is 65 s and the red light duration is 65 s andthe average waiting time of A1Queue is 1685.33 s. Simi-

larly, Table 2 presents that the average waiting times inD1Queue, E1Queue, and F1Queue are very high in the firstintersection. It also shows when the green light duration is65 s and the red light duration is 65 s, the average waitingtime of D1queue is 1656.2, which is an optimal solution

among these 6 cases (see Fig. 6). From the above descrip-tion, thus, in the following, we further fix the green lightduration to 65 s and each time increase the red light dura-tion by 15 s (i.e., g= 65, r = 50; g= 65, r = 65; g= 65,r = 80; g= 65, r = 95; g= 65, r = 110; and g= 65,r = 125 s) as shown in Figs. 7 and 8. Also, we fix the red

light duration to 65 s and each time increase the green lightduration by 15 s (i.e., g= 50, r = 65; g= 65, r = 65; g= 80,r = 65; g= 95, r = 65; g= 110, r = 65 and g= 125,r = 65 s) as shown in Figs. 9 and 10. Table 3 illustratesthe average waiting times in A1Queue are 1125.32,

D1Q

ueue

D2Q

ueue

D3Q

ueue

E3Queue

E2Queue

E1Queue

F1Queue

F2Queue

F3Queue

0

500

1000

1500

2000

r125g125

r110g110

r95g95

r80g80

r65g65

r50g50

Fig. 6. The average waiting times for the same light durations (Roads D, E, and F).

Table 2The average waiting times for the same light durations (roads D, E, and F)

D1Queue D2Queue D3Queue E1Queue E2Queue E3Queue F1Queue F2Queue F3Queue

r50g50 1683.6 32.3026 29.5532 1683.6 32.2915 29.5403 1683.6 32.3031 29.5391r65g65 1656.2 34.6355 34.6568 1656.2 34.6356 34.6853 1656.2 34.6323 34.6492r80g80 1709.95 32.1839 35.0228 1709.95 32.1829 35.0011 1709.95 32.1801 35.0147

r95g95 1693.3 34.8785 34.9261 1693.3 34.8729 34.9295 1693.3 34.8844 34.8996r110g110 1717.31 32.1179 32.1523 1717.31 32.1194 32.1666 1717.31 32.1207 32.1584r125g125 1685.4 34.5409 31.8162 1685.4 34.5437 31.8157 1685.4 34.5425 31.8154

0

400

800

1200

1600

2000

2400

2800

A1Q

ueue

A2Q

ueue

A3Q

ueue

B1Q

ueue

B2Q

ueue

B3Q

ueue

C1Q

ueue

C2Q

ueue

C3Q

ueue

r125g65

r110g65

r95g65

r80g65

r65g65

r50g65

Fig. 7. The average waiting times for different red light durations and same green light durations (A, B, and C).

D1Q

ueue

D3Q

ueue

D2Q

ueue

E1Queue

E2Queue

E3Queue

F1Queue

F2Queue

F3Queue

0

300

600

900

1200

1500

1800

2100

2400

r125g65

r110g65

r95g95

r80g65

r65g65

r50g65

Fig. 8. The average waiting times for different red light durations and same green light durations (D, E, and F).



9/12

1685.33, 2124.22, 2475.5, 2786, and 3006.5. We can knowwhen the red light duration is 50 s and the green light dura-tion is 65 s, the average waiting time ofA1Queue is 1125.32,which is the lower (see Fig. 7). Furthermore, Table 4 showsthat the average waiting times of A1Queue are 2233.75,1656.2, 1271.3, 900.13, 643.9, and 381.17. The averagewaiting time of 381.17 is the lowest, which means thatthe red light duration is 125 s and the green light durationis 65 s (see Fig. 8). Therefore, compare Tables 3 and 4, the

average waiting time of 381.17 is the minimum when thered light duration is 125 s and the green light duration is65 s for roads D, E, and F. However, from Table 5, we

can find an interesting situation, the average waiting timesin A1Queue are 2259.11, 1685.33, 1291.6, 911.29, 656.05,and 387.73. The average waiting time of 387.73 is the low-est when the red light duration is 65 s and the green lightduration is 125 s for roads A, B, and C (see Fig. 9). Table6 also shows the average waiting times of D1Queue are1110.2, 1656.2, 2097.05, 2461.29, 2746.94, and 3005.2respectively (see Fig. 10). This means when the red lightduration is 65 s and the green light duration is 125 s, the

average waiting time, 387.73 s, is the lowest. Actually, thisis because if the traffic light signals on roads A, B, and C inthe EastWest direction are green then those of roads D, E,

A1Q

ueue

A2Q

ueue

A3Q

ueue

B1Q

ueue

B2Q

ueue

B3Q

ueue

C1Q

ueue

C2Q

ueue

C3Q

ueue

0

300

600

900

1200

1500

1800

2100

2400

r65g125

r65g110

r65g95

r65g80

r65g65

r65g50

Fig. 9. The average waiting times for the same red light durations and different green light durations (A, B, and C).

D1Q

ueue

D3Q

ueue

D2Q

ueue

E1Queu

e

E2Queu

e

E3Queue

F1Queue

F2Queue

F3Queu

e

r65g125

r65g110

r65g95

r65g80

r65g65

r65g50

300

600

900

1200

1500

1800

2100

2400

2700

3000

0

Fig. 10. The average waiting times for the same red light durations and different green light durations (D, E, and F).

Table 3The average waiting times for different red light durations and same green light durations (A, B, and C)


r50 g65 1125.32 39.972 22.457 1125.32 39.9695 22.4688 1125.32 39.9627 22.4461r65 g65 1685.33 48.4127 29.0645 1685.33 48.4121 29.0378 1685.33 48.4125 39.056r80 g65 2124.22 57.0281 35.7112 2124.22 57.0258 35.725 2124.22 57.0285 35.7176r95 g65 2475.5 65.4119 42.22 2475.5 65.4136 42.1975 2475.5 65.4142 42.1887r110 g65 2786 73.8641 48.7588 2786 73.841 48.7539 2786 73.8721 48.7676r125 g65 3006.5 82.2903 55.2945 3006.5 82.3039 55.3223 3006.5 82.3243 55.3138

Table 4The average waiting times for different red light durations and same green light durations (D, E, and F)


r50g65 2233.75 41.2237 38.1059 2233.75 41.219 38.0922 2233.75 41.2204 38.1232r65g65 1656.2 34.6355 34.6568 1656.2 34.6356 34.6853 1656.2 34.6323 34.6492r80g65 1271.3 26.7653 29.3691 1271.3 26.7652 29.3687 1271.3 26.7623 29.3538r95g65 900.13 25.0126 25.0368 900.13 25.0112 25.0462 900.13 25.0116 25.0469r110g65 643.29 20.3486 20.3846 643.27 20.3434 20.3625 643.27 20.3472 20.381

r125g65 381.17 20.1356 17.9425 381.17 20.1355 17.9729 381.17 20.1356 17.956



10/12

and F in the NorthSouth direction are red. When the traf-fic duration for EastWest direction is set longer, in con-trast, that of NorthSouth direction will be shorter.Consequently, according to the above analysis, we sum-mary that if the red light duration is set to 65 s and thegreen light duration is set to 125 s, the average waiting timewill be the lowest, which means that it is the best solution.When the red light duration is set to 65 s and the green lightduration is set to 110, the average waiting time is the sec-ond lowest, which means that it is the second best solution.When the red light duration is set to 65 s and the green light

duration is set to 95, the average waiting time is the thirdlowest, which means that it is the third best solution. Thesesimulation results will be used for knowledge reasoning for

traffic improvement. At this circumstance, it does not seemworth to continually adjust the red/green light duration.Particularly, if we reduce the red light duration at an inter-section, it will decrease the green light duration in the otherside at the intersection, which will probably increase theaverage waiting time in the other side.

To find out the best traffic performance, we furtheradjust the interarrival time and departure time for a carpassing the stop line at an intersection. First, we assumethat the traffic on roads D, E, and F are very light so wechange the interarrival time on roads D, E, and F from

1.7 to 3.4 s. This case is represented as ABC_1.2 s, theinterdeparture time at the first intersection on roads A,B, and C has been modified from 1.2 s to 0.6 s (i.e., it

Table 5The average waiting times for the same red light durations and different green light durations (A, B, and C)


r65g50 2259.11 60.1125 47.5528 2259.11 60.1353 47.5378 2259.11 60.123 47.5485r65g65 1685.33 48.4127 29.0645 1685.33 48.4121 29.0378 1685.33 48.4128 29.056r65g80 1291.6 42.337 39.7984 1291.6 42.3542 39.7762 1291.6 42.3568 39.8164r65g95 911.29 37.3408 37.3982 911.29 37.3557 37.3734 911.29 37.3521 37.3913

r65g110 656.05 34.1988 31.9347 656.05 34.1983 31.9025 656.05 34.1993 31.9155r65g125 387.73 31.3646 31.3834 387.73 31.3609 31.38 387.73 31.3591 31.3799

Table 6The average waiting times for the same red light durations and different green light durations (D, E, and F)


r65g50 1110.2 27.5412 27.5616 1110.2 27.5413 27.5699 1110.2 27.544 27.5764r65g65 1656.2 34.6355 34.6568 1656.2 34.6356 34.6853 1656.2 34.6323 34.6492r65g80 2097.05 41.6674 41.6899 2097.05 41.6633 41.7105 2097.05 41.6664 41.698r65g95 2461.29 48.4843 48.5221 2461.29 48.4877 48.5158 2461.29 48.4906 48.5129r65g110 2749.94 56.1156 56.1541 2749.94 56.1148 56.1479 2749.94 56.1085 56.1693r65g125 3005.2 62.994 63.013 3005.2 63.0042 63.0129 3005.2 63.9992 63.0467

Table 7The average waiting times for red = 65 and green = 125 (A, B, and C)


ABC_1.2 387.73 31.3602 31.3882 387.73 31.3563 31.4052 387.73 31.3582 31.3926ABC1_0.6 292.04 156.48 29.2196 292.04 156.48 29.221 292.04 156.48 29.2149ABC12_0.6 199.01 147.92 156.48 199.01 147.92 156.48 199.01 147.92 156.48ABC_0.6 16.9542 12.1409 12.8234 16.9542 12.1384 12.8046 16.9542 12.1413 12.8249

A1Q

ueue

A2Q

ueue

A3Q

ueue

B1Q

ueue

B2Q

ueue

B3Q

ueue

C1Q

ueue

C2Q

ueue

C3Q

ueue

200

400

600

0

Fig. 11. The average waiting times for red = 65 and green = 125 (A,B, and C).



11/12

means that an extension lane is allowed so the averagespeed can be increased) and named it as ABC1_0.6. Then,that of the first and second intersection on roads A, B, andC is changed to 0.6 and represented as ABC12_0.6. Finally,all departure times at the three intersections on roads A, B,and C are changed and called as ABC_0.6. Based on Table7 and Fig. 11, we know that all models (e.g., ABC_1.2,ABC1_0.6, ABC12_0.6, and ABC_0.6) improve traffic per-formance, however, the model for ABC_0.6, which pre-sents all departure times at three intersections changed

from 1.2 to 0.6, is the best alternative because the averagewaiting times of A1Queue, B1Queue, C1Queue, A2Queue,B2Queue, C2Queue, A3Queue, B3Queue, and C3Queueare sharply dropped. Fig. 12 and Table 8 also showthe average waiting times of D1Queue, E1Queue, andF1Queue are greatly reduced and those of D2Queue,E2Queue, F2Queue, D3Queue, E3Queue, and F3Queueremain the same values as those in Table 6.

4. Conclusions and future work

This paper proposes a new framework for dynamic andautomatic traffic light control expert systems for improvingtraffic congestion problem. To know how to automaticallyand dynamically set the time duration of red (green) lightsignals, we also design a simulation model for improvingtraffic problem in rush hours. In order to analyze systemperformance, we design a traffic simulation model, whichconsists of six submodels. Each submodel represents a roadthat has three intersections. The simulation results physi-cally prove the efficiency of the traffic system in an urbanarea, because the average waiting time of cars at everyintersection is dropped down sharply when the red lightduration is 65 s and the green light time duration is 125 s.

Meanwhile, further analysis also shows if we keep the inter-

arrival time of roads A, B, and C, and change that of roadsD, E, and F from 1.7 to 3.4 s. Besides, four cases includingABC_1.2, ABC1_0.6, ABC12_0.6, and ABC_0.6 areadopted. The case of ABC_1.2 means that the interdepar-ture time at the three intersections is 1.2 s. The case ofABC1_0.6 presents that the interdeparture time at the firstintersection on roads A, B, and C is 0.6 s, and that of thesecond and third intersection on roads A, B, and C remains1.2 s. ABC12_0.6 describes that the interdeparture time atthe first and second intersections on roads A, B, and C is

0.6 and remain that of the third intersection is 1.2 s.Finally, ABC_0.6 stands for the interdeparture time atthe three intersections is 0.6. The result shows when theinterdeparture times at the three intersections on roadsA, B, and C are equal to 0.6 s, the total performance ofthe simulation model is the best.

Although this paper presents and analyzes the DATL-CES, there are still several aspects where we can furtherimprove its functions. In particular, we can extend the sim-ulation model to use two-way roads or allow cars turningleft or right to let the model more close to the reality. Inaddition, because we can collect traffic flow and average

car speed by using RFID technology, the method ofdynamically finding a best route or a second optimal routefor road navigation systems will be also a major researchissue in the future.

References

Chabini, I. (1997). A new algorithm for shortest paths in discrete dynamicnetworks. In 8th IFAC/IFIP/IFOrS symposium on transportationsystems, Tech Univ Crete, Greece (pp. 1618).

Chabini, I. (1998). Discrete dynamic shortest-path problems in transpor-tation applications. Transportation Research Record, 170175.

Chen, Y. L., & Yang, H. H. (2000). Minimization of travel time andweighted number of stops in a traffic-light network. Transportation

Research B, 34, 241253.

D1Q

ueue

D3Q

ueue

D2Q

ueue

E1Queue

E2Queue

E3Queue

F1Queue

F2Queue

F3Queue

0

2040

6080

100120140160

180

Fig. 12. The average waiting times for red = 65 and green = 125 (D, E, and F).

Table 8The average waiting times for red = 65 and green = 125 (D, E, and F)


ABC_1.2 152.6 70.7229 66.8913 152.6 70.7222 66.915 152.6 70.7217 66.8833ABC1_0.6 152.6 70.7286 66.897 152.6 70.7279 66.8934 152.6 70.7295 66.8736ABC12_0.6 152.6 70.7244 66.8862 152.6 70.7246 66.8973 152.6 70.727 66.8944

ABC_0.6 152.6 70.7269 66.8833 152.6 70.7267 66.8921 152.6 70.7191 66.8918



12/12

Chen, Y. L., & Yang, H. H. (2003). Minimization of travel time andweighted number of stops in a traffic-light network. European Journalof Operational Research, 144, 565580.

Eriksson, H. (1996). Expert systems as knowledge servers. IEEE, ExpertSystems and the Web, 1418.

Fay, A. (2000). A fuzzy knowledge-based system for railway trafficcontrol. Engineering Applications of Artificial Intelligence, 13, 719729.

Findle, N. V., Surender, S., & Catrava, S. (1997). On-line decision aboutpermitted/protected left-hand turns in distributed traffic signal control.

Engineering Applications Artificial Intelligence, 10(3), 315320.Grau, R., & Barcelo, J. (1992). An experience in demand-responsive traffic

control. In Proceeding of ist Meeting of the Euro Working Group in theUrban Traffic and Transportation, Landshut. Germany: TechniczlUniversity of Munich.

Hoyer, R., & Jumar, U. (1994). An advanced fuzzy controller for trafficlights. Annual Review in Automatic Programming, 19, 6772.

Ikeda, T., & Imai1, H. (1994). Fast a algorithms for multiple sequencealignment. IPSJ SIG Notes 94-AL-42-7, IPSJ.

Ikeda, T., Hsu, M.Y., & Imai, H., 1994. A fast A., Algorithm for findingbetter routes by AI search techiques. In: Vehicle navigation andinformation systems conference proceedings, IEEE (pp. 291296).

Liao, S. H. (2002). Problem solving and knowledge inertia. ExpertSystems with Applications, 2131.

Liu, B. (1997). Routing finding by using knowledge about the roadnetwork. IEEE Transactions on System, man, and Cybernetics-Part A:Systems and Humans, 27(4), 425430, July.

Maniccam, S. (2006). Adaptive decentralized congestion avoidance in two-dimensional traffic. Physica A, 343, 512526.

Messmer, A., & Papageorgiou, M. (1994). Automatic control methodsapplied to freeway network traffic. In 12 IFAC World Congress,Australia, vol. 9 (pp. 233238).

Nooralahiyan, A. Y., Dougherty, M., Mckeown, D., & Kirby, H. r.(1997). A field trail of acoustic signature analysis for vehicle classifi-cation. Transport Research C, 5(3/4), 165177.

Schaefer, L., Upchurch, J., & Ashur, S. A. (1998). An evaluation offreeway lane control signing using computer simulation. Mathematicaland Computer Modelling, 27(911), 177187.

Sheu, J. B. (2006). A composite traffic flow modeling approach forincident-responsive network traffic assignment. Physica A, 367,461478.

Stoilova, K., & Stoilov, T. (1998). Traffic noise and traffic light control.Transportation Research D, 3(6), 399417.

US Department of transportation (2007) Congestion Mitigation. .

Wangermann, J. P., & Stengel, R. F. (1998). Principled negotiationbetween intelligent agents: a model for air traffic management.

Artificial Intelligent in Engineering, 12, 177187.Wen, W., & Hsu, H. W. (2005). A route navigation system with a new

revised shortest-path routing algorithm and its performance evalua-tion. WIT Transactions on the Built Environment (Urban Transport),77, 733743.

Wen, W., & Yang, C. L. (2006). A dynamic and automatic traffic lightcontrol system for solving the road congestion problem. WITTransactions on the Built Environment (Urban Transport), 89, 307316.

Xia, L., & Shao, Y. (2005). Modelling of traffic flow and air pollutionemission with application to Hong Kong Island. EnvironmentalModelling & Software, 20, 11751188.

Yang, X., & Recker, W. (2005). Simulation studies of informationpropagation in a self-organizing distributed traffic information system.

Transportation Research Part C, 13, 370390.

http://www.fhwa.dot.gov/congestionmitigation/congestionmitigation.htmhttp://www.fhwa.dot.gov/congestionmitigation/congestionmitigation.htmhttp://www.fhwa.dot.gov/congestionmitigation/congestionmitigation.htmhttp://www.fhwa.dot.gov/congestionmitigation/congestionmitigation.htm

ec1-dynamic traffic light control system

Documents