# Traffic intersection control

### Chapter 1

### INTRODUCTION

The reason for the participation and taking up this area for this project is to provide a tool for the analysis of traffic intersection performance. The various topics included in the study of fuzzy logic are what does fuzzy basically mean, basic notion and concepts of fuzzy sets, fuzzy set operations such as union, intersection, complementation, triangular norms. A detailed case study is carried out on traffic intersection control and distributed traffic control. Factors considered in the study include detailed look into fuzzy logics and fuzzy sets, traffic intersection control, signal timing algorithm, size and shape of various intersections (junctions), as well as the traffic. With a discrete-event simulation, the effects of changing or varying any of one of these factors can be monitored easily with performance measures such as the average delay time at the intersection, the time taken to change various colors in the traffic signals, average waiting time in a queue.

The span of this project is restricted to four-way signal-controlled intersections, because it is most widely found in all the major countries in the world. The effect of a network of multiple intersections linked together is considered in this study later on. Initially the various intersections studied are assumed to be isolated from any neighboring intersections but in the later studies looks into Distributed Traffic Control. Reliable with this postulate that a car that reaches the front of its lane of traffic and has a green signal is understood to be able to exit the intersection with no problem. Exit traffic from the intersection is not modeled. Various problems in exiting of traffic like accident at the junction, traffic controlling policemen present, traffic conjunction due to festivals or important days of the year, etc, is not taken in consideration. At present the model considers that there is at the most one lane from which the vehicles can turn left and one lane from which the vehicles can turn right for each direction of traffic. Lanes from which vehicles can go straight ahead are assumed to be infinite in length. The numbers of cars that can fit in straight or right turn lanes are not limited by their size i.e. lanes from which cars can turn right are also assumed to be infinite in length. The model can be expanded to include more specialized intersections by common logic, but this is the current state of the simulation program. Also, the simulation code only incorporates constant signal timing plans. However, the code has been written in such a way that an actuated signal timing plan, for any specific intersection can be easily interpreted and implemented. Each year a large amount of money is spent in studying and trying to improve traffic intersections. There are various ways and methods to improve traffic on the roads which indirectly helps in building the country�s economy.

Originally, traffic intersection research paid attention on analytically describing vehicles' delay in terms of the intersection characteristics. One of the most prominent early works is Webster. For more recent analytical delay models and queuing theory approaches see Hurdle and Hagen and Courage. A discussion of more complex intersection models can be found in Meneguzzer. For a detailed study of intersection characteristics and performance, the flexibility of computer simulation is preferable to analytical approaches. Many privately and publicly funded simulation models exist for the purpose of studying traffic networks. A few examples are UTCS, TRANSYT, SCOOT, DITCS, RT-TRACS, and RHODES. Most of these programs are intended as a tool to aid in setting traffic signal timing algorithms for a network of intersections. One program developed exclusively for the purpose of evaluating traffic network management systems is MITSIM, described in Yang and Koutsopoulos.

The focus of this work differs from others in that this simulation model offers the user the chance for more detailed intersection analysis. Performance measures of an intersection's performance can be split between individual lanes of traffic, distinct turn patterns, or both. Additionally, the advantages and disadvantages of a number of intersection characteristics can be easily measured, including, size, number, and type of lanes, as well as the signal timing algorithm.

### Chapter 2

### BASIC NOTIONS AND CONCEPTS OF FUZZY SETS

### 2.1. History

Fuzzy sets and fuzzy logic is one of the most up-and-coming areas in the field of modern-day technologies of information and data processing. Fuzzy sets have a rather brief history. This is because it has come into existence only in 1965 by Prof. Lotfi A. Zadeh of University of California at Berkeley [1]. Fuzzy set theory involves capturing, operating and representing with linguistic notion-objects with unclear or vague boundaries. Borel and Lukasiewicz , two brief experts shed light on the essence of the difficulties and problems involving these eventual solutions with vague boundaries.

Modern studies spread across various areas, from pattern recognition, control, and knowledge based systems to computer vision and artificial intelligence. There are a significant number of direct real world implementations ranging from house hold appliances to industrial installations. All of these involve fuzzy logic at some level or the other. It may be at the most primary level or at the most important level of the system but it is used at some point or the other. Fuzzy set theory has innumerous applications in various fields- automata theory, artificial intelligence, control theory, computer science, decision making, medical diagnosis, neural networks, expert systems, robotics, pattern recognition, social sciences are a few of them. Apart from this, it is being applied on a majorly in industries for building machine (engines, cars, ships, turbines, etc.), controls (Sendai subway in Japan, etc.) and for military purposes.

Fuzzy set theory is a generalization and also an extension of crisp set theory (CST). Thus, the basic ideas and the theme of CST will be reflected in FST also.

### 2.2 Concept of Fuzzy Set

As mentioned previously fuzzy set is an extension of the crisp set theory. Just as in a crisp set on a universal set U which is defined by its characteristic function from U to {0,1}, on the other hand in a fuzzy set the universal set U is defined by its membership function from U to [0,1]. The main difference in the above two sentences is in the brackets. In the former one the only values that are permissible are 0 and 1 and nothing else, whereas in the latter one all the values between 0 and 1 are allowed. The values can be anything including 0 and 1. Therefore there are infinite numbers of values that come into the picture. This is one of the major differences between crisp set theory and fuzzy set theory. Due to the easiness of fuzzy sets there has been a fundamental change in the approach of a particular problem. Consider the following example,

What would you do in such a case? This is the beauty of fuzzy logic. This illustration shows us how we can utilize fuzzy logic. Fuzzy logic means that the input that are given to a system are vague or unclear. They need not be precise as in the case of this illustration.

Let us consider another example. If we measure the height of the students in a particular class, the input in this case is the height of the students. Using the classical method we assign a particular benchmark to say whether a student is tall or not. Say we consider height greater than or equal to 180cm as tall then a student who is 179cm tall is also considered short. So when we take it in a fuzzy sense we mark students greater than 180 as tall and lesser than 160 as short. The students whose height lies between this range of 160cm and 180cm are neither tall nor short.

### 2.3 Differences between Fuzzy and Crisp Sets

� Conventional Method ( Crisp )

� Complex system � difficulty in modeling/solving

� Precision is difficult to obtain

� Does not utilize the human reasoning which don�t have sharp boundaries.

� Fuzzy Logic

� Complex system � easy to represent/solve

� Approximate solution : but as good

� Allows inputs with imprecise nature to be specified

� Cost-effective way to model complex systems, giving qualitative description

Here are certain reasons why we prefer Fuzzy Logic over Crisp Logic.

� Fuzzy logic is basically understood with ease.

The mathematical concepts behind fuzzy analysis are very straightforward. Fuzzy logic is a more instinctive advance without the far-reaching difficulty.

� Fuzzy logic is very flexible.

With any known system, it is easier to layer on further functionality devoid of starting again from the entire scratch.

� Fuzzy logic is liberal of data which is imprecise.

Everything is imprecise if you look closely enough, but more than that, most things are imprecise even on careful inspection

� Fuzzy logic can form nonlinear functions of random difficulty.

We can generate a fuzzy system to match any set of input-output data. This procedure is made mainly easy by adaptive techniques like �Adaptive Neuro-Fuzzy Inference Systems�(ANFIS), which is available in Fuzzy Logic Toolbox.

� Fuzzy logic can be built on zenith of the knowledge of experts.

In straight distinction to neural networks, which take training data and produce opaque, dense models, fuzzy logic lets you rely on the knowledge of populace who already know and understand your system.

� Fuzzy logic can be easily blended with usual control techniques.

Fuzzy systems don't essentially replace old conventional control methods. In many cases fuzzy systems add to them and abridge their execution.

� Fuzzy logic is based on ordinary verbal communication.

The basis for fuzzy logic is the foundation for human communication. This examination underpins many of the other statements about fuzzy logic. Because fuzzy logic is built on the structures of qualitative explanation used in day to day language, fuzzy logic is much simpler to use.

The last statement is possibly the most significant one and deserves additional debate. Natural language, which is used by normal people on a daily basis, has been produced by thousands of years of human history to be suitable and efficient. Sentences written in regular language represent a success of efficient communication.

When should we not use Fuzzy Logic Approach

Fuzzy logic is definitely not a cure-all like most of the approaches to this field. When should you not use fuzzy logic? The safest declaration is the one made in the foreword: fuzzy logic is a suitable method to plot an i/p space to an o/p space. If there is a more convenient method to a given problem we better use that rather than going into for this fuzzy logic approach. Fuzzy logic is the codification of universal common intelligence � employs common sense when you put into practice and you will probably make the correct choice. Many controllers, for example, do a great work without the use of fuzzy logic. If we take the time to become proverbial with fuzzy logic, we will see that it can be a very commanding tool for dealing swiftly and resourcefully with vagueness and nonlinearity.

### 2.4 Common Control Application

One of the most basic and simple way to understand what fuzzy logic means is to use the example of cart-pole problem. This is also called the inverted pendulum problem due to the shape. Sometimes we see how a juggler balances a pole in the palm of his hand. This balancing act involves the movement of the palm in forward and backward direction and even left and right. But here we just consider forward and backward movement. This act of controlling the pole can be converted into a laboratory experiment usually called the cart-pole or inverted pendulum problem. The cart is a wooden block and a pole is attached in the middle of the top surface of the cart. This arrangement in the below:

The main objective of the task is to keep the pole in the vertical position or at least near to it. Now if we consider the conventional method it is really tedious to solve this problems because there are various equations governing this situation:

I (d2 ?/dt2) = VLsin ? - HLcos ?

V - mg = -mL[(d2 ?/dt2)sin ?+(d ?/dt)2cos ?]

These are obviously differential equations and will be simulating our system. This will be really tedious, difficult, error prone, as well as time consuming.

Now looking into fuzzy logic approach.

The inputs are:

? : +ve(P) right of the vertical line & -ve (N) left of the vertical line

x=d?/dt : +ve(P) pole is falling to the right & -ve (N) pole is falling to the left

u: +ve(P) direction is right & -ve (N) direction is to the left.

INPUTS ?: N Z P

x: N Z P

OUTPUT u: NB N Z P PB

Where P: +ve, N: -ve, Z: Zero, NB: -ve Big, PB: +ve Big

Now consider the FUZZY RULE BASE as given below

X

When ? = N and x = N then u = NB

Therefore we don�t have a particular fixed value for ? and x to gt the required answer. So the problem now becomes pretty simple. When ? is negative and x is also negative then the u has to be a big negative number.

### Chapter 3

### MATLAB

### 3.1 Basics

MATLAB is a high-performance language for technical computing. It inte- grates computation, visualization, and programming in an easy-to-use environ- ment where problems and solutions are expressed in familiar mathematical notation.

### Typical uses include the following:

� Math and computation

� Algorithm development

� Data acquisition

� Modeling, simulation, and prototyping

� Data analysis, exploration, and visualization

� Scientific and engineering graphics

� Application development, including building graphical user interfaces.

MATLAB is an interactive system whose basic data element is a matrix. This allows formulating solutions to many technical computing problems, especially those involving matrix representations, in a fraction of the time it would take to write a program in a scalar non-interactive language such as C.

The name MATLAB stands for Matrix Laboratory. MATLAB was written originally to provide easy access to matrix and linear algebra software that previously required writing FORTRAN programs to use. Today, MATLAB incorporates state of the art numerical computation software that is highly optimized for modern processors and memory architectures.

MATLAB has various functions to work with images. One of the most powerful tools in MATLAB is M- Function Programming. It is flexible and allows image manipulations.

### 3.2 Some Function References (Taken From MATLAB 7.7.0)

GUI Tools and Plotting anfisedit

Opening an ANFIS Editor Graphic User Interface

locatecluster

Interactive clustering GUI for fuzzy c-means and subclustering

fuzzy

Unlocks fundamental Fuzzy Inference System editor

memberfunctionedit

Membership function editor

plotfis

Used to plot FIS

plotmf Plans all membership functions for given variable

ruleedit

Rule editor and parser

ruleview

Rule viewer and fuzzy inference diagram

surfview

Open Output Surface Viewer

Membership Functions

dsigmf

Built-in membership function composed of difference between two sigmoidal membership functions

gauss2mf

Gaussian combination membership function

gaussmf

Gaussian curve built-in membership function

gbellmf

Generalized bell-shaped built-in membership function

pimf

?-shaped built-in membership function

psigmf

Built-in membership function composed of product of two sigmoidally shaped membership functions

sigmf

Sigmoidally shaped built-in membership function

smf

S-shaped built-in membership function

trapmf

Trapezoidal-shaped built-in membership function

trimf

Triangular-shaped built-in membership function

zmf

Z-shaped built-in membership function

FIS Data Structure

addmf

Adds an mf to an FIS

addrule

Adds a rule to FIS

addvar

Adds a variable to FIS

defuzz

Helps in defuzzifying a membership function

evalfis

Execute FI calculations

evalmf

Generic membership function evaluation

gensurf

Produces FIS output surface

getfis

System properties

mf2mf

Translate parameters between membership functions

newfis

Create new Fuzzy Inference System

parsrule

Parse fuzzy rules

readfis

Loads a FIS from file

rmmf

Eliminates membership function from FIS

rmvar

Eliminates variables from FIS

setfis

Set fuzzy system properties

showfis

Exhibits annotated FIS

showrule

Exhibits FIS rules

writefis

Save FIS to file

### 3.3 MEMEBERSHIP FUNCTIONS

There are many types of membership functions which are as follows:

Triangular, trapezoidal, Gaussian, bell shaped, s-shaped, z-shaped etc. the next sections discusses the steps in order to generate a membership function.

### 3.4 GENERATING MEMBERSHIP FUNCTIONS

The method proposed by Hong, et al. [3] will automatically generate a membership functions. This process partitions a set of information into classes that can be used to obtain membership functions. The process of the algorithm has a number of chief stepladders and consists of clustering the facts and information into classes. The membership functions are then generated from the classes obtained. The algorithm is described for one parameter, and is detailed below.

Step 1. Given a information set, there are n exercise samples. The values for the parameter in query, , X = x1,x2��xn are sorted into increasing order, denoted as Y = sort[X]=y1,y2��..yn. The values are sorted in order to determine an connection amid neighboring or adjoining values.

Step 2. The divergence amid adjoining values in the sorted data is determined. The difference obtained will supply a method to compute the connection between adjoining values. The disparity for a set of training set data is given by:

diffi = yi+1 - yi for i=(1,2,3,�n-1),

where yi and yi+1 are adjoining values in the data that is sorted.

Step 3. Find the similarities among adjacent values. The formula given below determines the similarities flanked by bordering values and maps them into real numbers in the range 0 and 1.

si = 1 - diffi / (C*ss) for - diffi < (C*ss)

0 otherwise

where

diffi is the distinction between bordering information,

ss is the standard deviation of diffi, and

C is the parameter of control.

The control parameter is used to decide the profile of the membership function.

Step 4. The statistics is grouped on the basis of their similarities. A threshold value, a, divides adjoining values into classes. The quantity of classes influenzes the number of membership functions. Determining the number of classes can be put together by a rule: If the similarity is superior or larger than the determined threshold value a, then the two adjoining data fit in to the similar class, otherwise the values are separated into dissimilar classes. That is, a fresh class is produced. Expressed as a formula,

IF(si>a)THEN yi,yi+1 e Ci, ELSE yi e Ci, yi+1 e Ci+1

where Ci and Ci+1 denote two different classes for the identical input or output constraint.

Step 5. The membership function for each and every class is definite. There are many different types of membership functions these include triangular, trapezoidal, and Gaussian, to name a few. These are the most commonly used membership functions. One of the most basic membership functions is the triangular membership function, and will be used for the remaining equations. The triangular membership function for class j consists of three different points: the central vertex point, bj, this is the maximum value of the membership function and the two endpoints, aj and cj. The central vertex point is determined for every class and is determined by the below given expression:

bj = (yi * si + yi+1 * ����.. yk * sk-1) / (si + (si + si+1)/2 + ���. sk-1)

where

j represents the jth class,

k represents the ending data index for this class,

i.e., data yi through yk fall into class j, and

si is the similarity between yi and yi+1.

The endpoints of the membership function, aj and cj, are determined by the use of interpolation. The following equations determine the right and left endpoints.

aj = bj � [ (bj-yi) / 1-�j (yi) ]

cj = bj � [ (yk-bj) / 1-�j (yk) ]

Where aj is the left endpoint and cj is the right endpoint. mj(yi) and mj(yk) is the membership determined by the formula:

�j(yi) = �j(yk) = min(si, si+1, �.. sk-1)

where k represents the maximum data index value within the class.

### 3.5 Basic Types of Membership Functions

The fuzzy logic toolbox has 11 in-built functions of membership types. All of the 11 functions are build from a number of essential functions: bitwise or piecewise functions which is linear in nature, the Gaussian distribution function, the signum curvature. They can even be quadratic or cubical in nature. For comprehensive information on any of the membership functions mentioned, there are several information given in the �Function Reference� of Fuzzy Logic. By principle, all membership functions have the letter mf at the conclusion of their names.

The most basic and simple membership functions are shaped by means of straight lines. Of these, the simplest being the triangular membership function, and it has the function name �TRIMF�. It is nothing more than a compilation of three points forming the shape of a triangle. One point representing the apex or the maximum value of the function and other two points representing the end points. The trapezoidal membership function, �TRAPMF�, has a level top and actually is just a abridged or truncated triangle curve. These straight line membership functions have the benefit of straightforwardness.

The generalized bell membership function is precised by three parameters and has the function name �GBELLMF�. The bell membership function has an extra parameter when compared to the Gaussian membership function, so it can move toward a non-fuzzy set if the liberated parameter is tuned properly. Because of their effortlessness and concise notation, Gaussian and bell membership functions are accepted methods for defining fuzzy sets. Both of these curves have the lead of being smooth and non-zero at all the possible points.

Even though the Gaussian membership functions and bell membership functions accomplish smoothness, they are not capable to identify asymmetric membership functions, which are significant in certain applications. Next we define the sigmoidal membership function, which is either open on the left or right. Asymmetric and closed i.e. not open to the left or right membership functions can be generated using two sigmoidal functions, so in addition to the basic �SIGMF�, we also have the divergence between two sigmoidal functions, �DSIGMF�, as well as the product of two sigmoidal functions �PSIGMF�.

Polynomial based curves report for more than a few of the membership functions in the toolbox. Three related membership functions are the Z, S, and Pi curves. These functions are named on the basis of their shapes. The function �ZMF� is the asymmetrical polynomial curve which is open to the left, �SMF� is the spitting image function that opens to the right hand side, and �PIMF� is zero on both of the extremes with an elevation in the centre.

### 3.6 SUGENO FIM (Fuzzy Inference Method)

The fuzzy inference process taken into consideration so far is the Mamdani�s FIM, the most common line of attack. This division highlights the Sugeno, technique of fuzzy inference. Introduced in 1985, it is comparable to the Mamdani manner in various factors. The first two parts of the Sugeno�s fuzzy inference process, i.e. inputs being fuzzified and even the application of the operator, are accurately the same. The main dissimilarity between these two FIM i.e. Mamdani and Sugeno is that the latter method yield membership functions are moreover linear or constant.

A most common rule in a Sugeno fuzzy model has the expression:

If Input 1 = a and Input 2 = b, then Output is c = xa + yb + z

For Sugeno model of a zero order, the output point c is fixed (x=y =0).

The output point ci of every rule is prejudiced by the strength hi. Let us consider an example, for an �AND� rule with Input 1 = a and Input 2 = b, the firing strength is

where F1,2 (.) are the function of membership for the two inputs 1 and 2.

The concluding output of the system is the biased standard of all rule outputs, calculated as

where N represents the number of rules.

A Sugeno rule functions as shown in the diagram given below:

The first shows the fuzzy tipping model developed, which a very common example adapted for use as a Sugeno system. It is very commonly the fact that singleton o/p functions are totally enough for the wants of a given problem.

a = readfis('tippersg');

gensurf(a)

The best way to view order-1 Sugeno base is to consider of every rule as determining the position of a affecting singleton. That is, the singleton o/p spikes can shift around in a linear manner in the o/p region, which depends on what the i/p is. This in-turn tends to make the base information very packed together, resourceful and proficient. Higher-order Sugeno fuzzy models are probable, but they bring in notable complication with little obvious advantage. Sugeno fuzzy models whose yield mfs are superior than order 1 are not supported by FLT software.

Due of the linear reliance of every rule on the i/p variables, the method of Sugeno is perfect for performing as an interpolating supervisor of many linear controllers that are to be functional, to different working circumstances of a dynamic nonlinear structure. For example, the functioning of an aircraft may change spectacularly with elevation and Mach number. Linear controllers, though straightforward to compute and well suited to any specified flight condition, must be restructured frequently and easily to keep up with the varying condition of the flight medium. A fuzzy Sugeno conclusion system is tremendously well suited to the mission of effortlessly interpolating the gains which are linear that would be functional across the input space; it is a expected and efficient gain scheduler. Correspondingly, a Sugeno system is suitable for modeling nonlinear systems by interpolating among numerous linear models.

### Chapter 4

### TRAFFIC INTERSECTION CONTROL

Most urban traffic network links crisscross recurrently, leading to a variety of conflicts between the flows in the traffic. This function of intersection is often a significant factor in shaping the overall competence and performance of a network, therefore traffic engineers continuously face the problem of controlling and checking flow at intersections to improve the performance.

### 4.1 Three Phases of Traffic Lights

Traffic lights are used to control car flowing through most of the cities intersections. The cycle length of a signal is the time period required for one complete sequence of signals at a given intersection. The signal light is basically divided into three different part:

1) Red Phase

2) Amber Phase

3) Green Phase

The cycle length is normally divided into a number of phases, each phase being a part of the time cycle allocated to one or more traffic and pedestrian movements. The green phase is a particular phase provides a green light ( right of way ) for a given particular direction. Green time represents the time period or the amount of tie for which the green phase is operative. Similarly we have for the other two phases as well.

There are various things that have to be kept in mind while designing these phases for a particular traffic signal. This may differ according to the location, the density of the vehicles, the density of the pedestrians, no of lanes available, reaction time for the vehicle users to react while waiting in the queue, etc. These are some points that have to be kept in mind while designing the time period for the three phases. In order for proper control and free flow of the traffic.

### 4.2 Planning and Strategy

Often there are many plans and strategies that are employed by the traffic engineers in order to keep the traffic flow under check. These consist of changing the green time ( and consequently changing the cycle length ) as a function of the incoming traffic so that the cars share the intersection more effectively and efficiently. But there is a problem that arises; each intersection has its own specific characteristic features ad physical layout, rate at which traffic flows in, turning movements, density of the pedestrians, and so on. There are mathematical models describing the traffic flow though an intersection for a given flow density, but it is really difficult to take into account the fluctuations of the traffic that flows towards a junction from different directions. As the flow of the traffic increases in a particular direction the vehicles are forced to wait in the queue for a longer time and this increases the load on the traffic that is going to arrive at the junction after some time.

Clearly, the state of an intersection may be clearly characterized by the number of vehicles arrivals, length of the queue and the control decision is the green phase or the green time. There are other parameters like the physical layout of the intersection, number of lanes per approach and the dimensions of the lane.

### 4.3 Policies in traffic control systems

There are mainly two policies under which traffic control is classified, these are:

1) Fixed Time Systems

2) On line Systems

### 4.3.1 Fixed Time Systems

In this type of policy or plan the various traffic plans are applied on-line but they are generated off-line.it consists of a control plan which is computed from average measured traffic flows. The controller that is used in this type of policy can store a number of such plans and then they can be implemented during different hours of the day as per the use. This particular policy requires periodic check on the traffic at each and every signal separately. Therefore surveys have to be conducted to keep a record of the average traffic flow at different junctions.

### 4.3.2 On-line Systems

In this policy the various plans are generated on-line and are then implemented directly for the traffic control. For example, during office hours like around 9 a.m. the green phase will be on for a longer period of time when compared to the time its on for at around 12 noon. Also when during late night when the number of pedestrians are very few compared to the number of car, the green phase will be extremely long. In such cases we also use certain type of prompters that are used by the pedestrians and the signal turns red for the vehicles only when the traffic controller is prompted to do so. Otherwise it remains in the green mode.

Both these policies have their own advantages as well as disadvantages but their common aim is to minimize the average vehicle delay intersections cause and decreases the length and waiting time of the vehicles in the queue.

### 4.5 Passive control

When the amount of travel is fewer, no unambiguous control is necessary we use passive control. In this case the street users are requisite to comply with the fundamental rules and regulations of the road. Passive control like traffic symbols and signs, markings on the road etc. are used to balance the junction control. Some of the junction manage that are classified beneath passive control are stated below:

1. No organizing is required if the traffic approaching to a junction is really low for example during very early morning hours from 2 a.m. to 5 a.m., then by implementing the vital regulations of the road like driver on the left side of the street must give way and that from end to end arrangements will have precedence over turning movements, wearing a seat belt,etc. The driver is anticipated to comply with these necessary set of laws of the road.

2. Traffic signs: With the assist of caution signs, direct signs etc. it is able to make available some intensity of manage at an junction. Firstly give way , then two-way stop control, and finally all-way stop organize are a few of the basic examples.

The GIVE WAY control instructs the driver in the smaller road or the less important road to slow down to a bare minimum velocity and permit the car on the main road to carry on. Two way stop control instructs the car drivers on the less important streets should see into that the major conflicts are kept at bay. Finally an all-way stop control is regularly used when it is complicated to make a distinction between the most important and inconsequential roads in an meeting point. In such a case, STOP warning is positioned on all the approaches to the connection and the vehicle driver on all the approaches is obligatory to stop the car or bring it to the bare minimum check whether his way is clear and only then can he proceed. An important point to be kept in mind is that always the right lane gives priority to the left lane. The traffic control at 'at-grade' junction may be unrestrained in case of little traffic. Here it is essential for the street users to obey the fundamental laws of the street. Passive control for example traffic ciphers, road markings etc. are used to harmonize the crossroads control.

3. Traffic signs plus marking: In accumulation to the traffic sign boards, street markings also balance the traffic organization at junctions and intersections. Some of the examples contain hault line marking in order to remind the driver to stop at certain places before he proceeds, give way lines for reminding the driver to give ways to other more important lanes before he proceeds, arrow marking for directions, yellow line marking for showing no access to lane, etc.

### 4.6 Partial Control

In semi-control or partial-management, the vehicle users are gently guided to steer clear of conflicts. The two most important examples of this type of control are channelization and traffic rotaries.

Channelization: The traffic is alienated to flow all the way through specific paths by raising a segment of the street in the heart more often than not called as islands eminent by road markings. The conflicts in traffic activities are condensed to a huge extent in such cases. As it the suggests, in channelization, the transfer is intended to run through dissimilar channels and this bodily partition is made possible with the assist of certain blocks in the street like traffic islands, highway markings etc.

Traffic rotaries: It is a form of junction control within which the traffic is finished to run the length of one course around a traffic island. The important principle of this control is to exchange all the harsh conflicts like from end to end and right turn conflicts into softer problems like assimilation, weaving and diverging. This method is a form of `at-grade' intersection which is discussed in detail later is laid out for the progress of traffic such that no all the way through conflicts are there. Free-right turn is acceptable where as through traffic and left-turn traffic is required to move around the innermost island in \ direction which is clockwise in an arranged approach. Assimilation, weaving and diverging operations reduce to a great extend the conflicting activities at the rotary.

### 4.7 Active control

In this method strict implications are made on the drivers by the traffic control of the country like in the U.A.E it is done by the Road Transport and Authority (RTA).The driver cannot break any of the rules and regulations of such method on his own wish or else there is a heavy penalty or fine imposed on him. The two most important types are as follows:

1) Traffic Signals

2) Grade separated intersections

### Traffic signals

The most important basic rule that a traffic signal follows is time distribution approach. At any given occasion, with the facility of suitable signals, certain traffic activities are controlled where as certain additional arrangements are permitted to pass through the meeting point. There are different phases provided according to the traffic conditions but in most of the cases it is restricted to two or three. When the number of vehicles traversing the junction is very bulky, then this method of traffic signals is of utmost importance. The phase provided for the signal depends on the need. It can be two or more. If more than 2 of the phases are used, then it is termed as multiphase signal.

These signals present can operate on many modes. The most common modes that are used in most of the countries are permanent time signals as well as vehicle actuated signal. In the former one the the timing of all the phases that is green, amber and red is fixed prior to the manufacture of the signal. A survey is made in order to determine the intensity of the traffic at the junction and then these phase timings are calculated based on various parameter by the officials of the authority. In the vehicle actuated signals there are vehicle detectors are placed in the streets in order to determine the presence of vehicles. The functioning of such detectors is discussed later. When a vehicle is detected the green phase is actuated. This mode is more efficient but the drawback is that it is more complex to construct and is also costly.

### Grade separated intersections

There are basically two types of intersections they are a) at-grade intersections and b) grade-separated intersections. In the former intersection, each and every roadways unite or pass at the same perpendicular level. Grade alienated junctions allows the traffic to cross at diverse perpendicular levels. From time to time the geography may be supportive in setting up such an intersection. If this is not taken into thought then it can be still be constructed but the initial construction cost requisite will be extremely high. Therefore, they are mostly constructed on high velocity services like expressways, freeways etc. This type of method helps in increasing the capacity of the road because in these cases the traffic flow at extremely high speed and hence increases the capacity.

### Chapter 5

### FUZZY TRAFFIC CONTROLLER

The ground-breaking work in using these kinds of fuzzy set in the field of traffic control, by Pappis and Mamdani (197), considers a solitary junction only of two one-way streets. The outcome found by these by realization of fuzzy logic controller were compared to those obtained using the conservative, effective vehicle actuated controllers. Pappis and Mamdani considered the typical delay of the vehicles as the performance factor or in other words taking the delay of the vehicles as the output found out that the results obtained using fuzzy logic controller are better than those attained using the conventional actuated controllers.

The 3408319 depicts a block diagram for a fuzzy traffic controller (FTC) discussed here and its main futures are outlined:

1) Sensing Device

a) A set of two inductive loops which are separated by a distance �d�(one set per lane)

b) This helps us not only in vehicle detection but also in speed estimation.

2) Estimator

a) It computes each vehicles speed and the time it takes to cross the intersection, most importantly at the end of the green phase.

b) It provides us with the estimates of the arraival rate and the queue length at each approach.

3) Fuzzy Logic Controller

a) Determines the green phase in accordance to the traffic situation, more specifically it determines the green time.

4) State Machine

a) It controls the sequence of states the FTC should pass through

5) Adaptive Module

a) Changes the fuzzy controller settings to adjust its performance.

6) Traffic Light Interface

a) Provides the circuitry for visual display that is for turning on and off the lights in accordance to the fuzzy controllers decision.

### Chapter 6

### WHAT IS THE BASIC WORKING PRINCIPLE OF A SENSOR?

There is something unusual about the traffic lights that it detects the presence of a car. This section shows the basic working principle of such a sensor that help the traffic signal to detect the presence of the number of cars waiting in the queue at a red signal.

Some lights do not have any kind of detectors. For instance, in a large city, the traffic lights just function on what time of the day it is no matter what, there is always going to be traffic. In the outer edge and on nation roads, however, it is more common to find traffic signals working on the basis of detectors. These detectors help in many ways. They detect whether any car is in attendance at the junction, when loads of vehicles are stacked up in queue at an intersection. This helps to control the length of the green phase, or when cars make an entry at turn lane in order to make active the arrow light.

There are a number of technologies that are used for detecting the presence of a car, everything from lasers to air filled rubber hoses. By far the most common and simplest technique is the basic inductive loop. An inductive loop is a simple coil of wire entrenched in the surface of the streets. To install these inductive loops, they lay asphalt and then cut a channel in the asphalt with bulky machinery. The inductive wire is then placed in the channel and sealed with a rubbery amalgam. We can often see these big square or rectangular shaped loops cut in the asphalt road because this is where they place the detectors to determine whether a car is present or not.

Inductive loops work by detecting a change in the inductance of the atmosphere. To understand this basic process, let us first look into what exactly is inductance. The given below helps us in understanding this phenomenon

Given here is a simple DC source i.e. battery, a bulb, a wire of coils around a piece of core which is usually a bar of iron, and finally a switch. The coil of wire acts as an inductor.

If we take the inductor i.e. the piece of iron, out of the given circuit, then what we have here is a normal flashlight and a normal bulb which glows as long as the battery is connected to it and if we close the switch. When we close the switch the bulb glows. But due to presence of the inductor in the circuit shown above, the performance is astonishingly different. The light bulb acts as a resistor (the resistance within brings up warmth to make the fiber or filament in the bulb to brighten up). The coil in turn has much lower resistance because it is just a normal wire wound, so what we expect when we close the switch, bulb glows very dimly. This is because most of the current flows through the low-resistance path present in the circle. What happens as an alternative is that when we shut the switch, the bulb burns intensely and then becomes dimmer. When you unlock the toggle, the tuber burns very brilliantly and then dies out hastily.

The reason for this different behavior is due to the presence of the inductor in the track. When current originally begins to run in the loop, the loop wants to construct up a magnetic field of its own. While the field is under construction, the coil counters or inhibits the run of current through it. Once the field is set up, then current can run usually through the line. When the switch unlocks, the field about the loop keeps current running in the loop until the field falls down. The presence of this current keeps the bulb lit for short phase of instance even though the toggle is open.

The capacity of an inductor is mainly controlled by two important factors:

1) Number of coils.

2) The substance that the loop is enfolded around. This is also known as the core of the inductor.

. Each and every core has a different inductance. The inductance of a magnetic material is more than that of a non-magnetic material. There are also devices that can gauge the inductance of a coil. The typical unit of gauge is the Henry.

Let us say we take a coil of wire perhaps 4 feet in diameter, containing four or five loops of wire. We cut some grooves in the asphalt and place these coil in the grooves. We attach an inductance meter (an instrument that gauges the inductance) to the loop and see what the measure of inductance of the coil is. Now when we park a car over this coil and check the inductance again the inductance shown in the inductance meter will be much higher than shown before in the presence of no car. This increase in the inductance is due to the presence of a magnetic material (steel body of the car) in between the loops. The vehicle parked above the loop is stand-in like the heart of the inductor, and due to this there is a change in the inductance of the coil.

This is how the traffic lights uses these sensors in order to detect the presence of any vehicle. It continually checks the inductance of the coil in the street, and when there is a change in the inductance, we come to know that there is a car waiting. Basically this is how a sensor performs its job. There are several other methods to detect the presence of a car which are not discussed in this project.

### Chapter 7

### CASE STUDY

In this chapter we will look into different cases in order to determine the weight of the factors on which the green phase timing depends on.

1) No of cars waiting in queue �VS� green phase timing

2) No of pedestrians waiting �VS� green phase timing

3) No of lanes �VS� green phase timing

4) No of cars waiting in queue and pedestrians waiting �VS� green pahse timing

5) No of cars waiting in queue and pedestrians waiting �VS� green phase timing

(DETAILED)

6) No of cars waiting in queue and lanes �VS� green phase timing

7) No of cars waiting in queue, pedestrians waiting and lanes �VS� green phase timing

### Chapter 6

### CONCLUSION

In this report the topics that are covered are firstly what exactly is a fuzzy set. What do you mean by fuzzy, what is the difference between the conventional method and the fuzzy method. It highlights its uses and applications in other fields like automata theory, robotics, neural networks, medical diagnosis and in military uses. The report looks into the problem that most of the over populated cities is currently facing that is of traffic conjunction and how can fuzzy logic be used to solve this problem. Its also explains the working of a basic inductive detector, then the different types active and passive types of controls necessary to keep a check on traffic at various junctions. It looks into the functioning of MATLAB program and the subtopics that are considered in MATLAB is the fuzzy logic toolbox. By the use of Graphic User Interface the problem of traffic intersection control was formulated. Different membership functions are taken into considerations for different factor on which the green phase of the traffic lights depends like the number of cars waiting in the queue, number of pedestrians waiting, number of lanes, etc. It contains an in-depth study of traffic intersection control, fuzzy traffic control and its various components in the form of a block diagram. Various screenshots of FIS Editor, membership function editor for input and output, rule editor, rule viewer and finally the surface view to get a graphical representation of the solution which are all derived from MATLAB 7.0.0. (R2008b). There are two points that are taken under consideration while formulating the fuzzy logic problem. The following results are attained:

1) The green phase time increases as the no. of cars waiting in the queue increases. This means that more the no. of cars waiting in the queue at a given signal the time period for which the green phase is ON should increase so that more no. of cars pass through and hence decrease the traffic conjunction.

2) The green phase time decreases as the no. of pedestrians waiting at the signal increases. This means that more the no. of pedestrians the green phase should decrease implying that the red phase increases so that more no. of pedestrians can cross the road.

3) The green phase time increases as the no of lanes present decreases. This means that more the no. of lanes less the green phase timing.

### REFERENCES

1. C. Pappis and Mamdani, �A Fuzzy Logic controller for a Traffic Control System,for Traffic Junction� , IEEE Trans. On systems, Man, and cybernetics, vol. SMC-7. Oct 1977.

2. N. Findler and J. Stapp, �Distributed Approach to Optimized Control of Street Traffic Signals,� Jour. Of transportation Engineering, Vol 118, Jan/Feb 1992.

3. W.Pedrycz and F. Gomide, �An Introduction to Fuzzy Sets Analysis and Design�,Prentice Hall of India Private Ltd.

4. M. Ganesh, �Introduction to Fuzzy Sets and Fuzzy Logic�, Jay Print Pack Private Ltd, New Delhi

5. http://www.mathworks.co.uk/matlabcentral/fileexchange/11078

6. http://www.curiousread.com/2008/01/its-go-time-how-do-traffic-light.html

7. http://www.thefreelibrary.com/Adopting+A+Business-Oriented+Approach+To+Bandwidth+Management-a061620899

8. http://www.scribd.com/doc/17117473/What-is-Fuzzy-Logic

9. http://www.scribd.com/doc/18676927/Matlab-Toolbox

10. http://www.civil.iitb.ac.in/~vmtom/1107_nptel/ceTEI/39-Ltexhtml/p3/p.html