FUZZY- BASED COLLISION AVOIDANCE SYSTEM FOR AUTONOUMOUS DRIVING IN COMPLICATED

(1)

FUZZY- BASED COLLISION AVOIDANCE SYSTEM FOR AUTONOUMOUS DRIVING IN COMPLICATED

SCENARIOS

BY

ALMUTAIRI SALEH RASHED

A dissertation submitted in fulfilment of the requirement for the degree of Master of Science )Automotive Engineering(

Kulliyyah of Engineering

International Islamic University Malaysia

2018

(2)

ii

ABSTRACT

The overall safety of human beings and the avoidance of imminent accidents on roads was the primary motivation for this research. Accidents on roads have caused many incidents resulting to death, which refers to human mistakes after avoiding natural disasters that could occur on roads. We also can classify human mistakes into mental mistakes, which includes driving under the influence of alcohol, psychological problems or mental retardation. The second classification comprises the mistakes that could occur from the humans when they are in a healthy state of mentality. These accidents occur when the driver is not fully focused on the road, or there are obstacles in blind areas which are difficult to detect, due to fog or any other natural causes. The first step to assist the driver is to create a fully automatic system to avoid collisions, which is called a collision avoidance system (CAS). In this study, we present a fuzzy- based control approach for smart and safe obstacle avoidance in complicated traffic scenario, where there are static and dynamic obstacles (e.g. broken-down vehicles, wrong parking road-side vehicles, or moving vehicles, etc.). The fuzzy system takes an optimal decision to control the car throttle, braking, and steering to avoid collision using the available information on the road map (i.e. the distance to obstacles, the current traffic in the neighbouring lanes, the velocity of the front and rear car, etc.). Simulation results from three different scenarios involving a combination of dynamic and static or broken-down vehicles demonstrate that the fuzzy controlled car can effectively avoid obstacles or collisions in complicated traffic situations.

(3)

iii

ثحبلا ةصلاخ

ABSTRACT IN ARABIC

لمعل يسيئرلا عفادلا نإ لإل ليلاا بنجتلا( في ثبح

ص ط تابكرلما في )ماد لإا ةملاس وه

في قدلمحا رطلخا داعبإو ناسن

ت ثداولحا نا ثيح ،تاقرطلا ت

لإا لتق في ببس لىا بلاغلا في عجري كلذو ناسن

أ ثراوكلا دعبتستام دعب ينقئاسلا ءاطخ

ءاطخا مسقن نأ انعطتسا اذإو.هيعيبطلا يرثتا تتح ثدتح تيلا ءاطخلاا ىلع يوتتح يهو هيلقع ءاطخأ لىا ناسنلإا

لأا وا لوحكلا يلقعلا فلختلا وا هيسفنلا تاداهج

. لأا مسقلاو ةلق ببسب ميلس لقعلاو ثدتح تيلا ءاطخلاا يهو رخ

يؤرلا ةبوعص وا زيكترلا ة

ع هردقلا مدع وا ىل

أ بم نوكت ثيبح ءايشلاا فاشتك ن

ئاسلا نع ءايمع هقط يذلا مسقلا وهو ق

ع قئاسلا دعاسيام لوأ نإ.ثحبلا اذه هيوتحيس إ ىل

لأا داعب ص ط ماد آ وه ايل إ وتح ى ماظن ىلع هبكرلما آ

لي عيطتسي لماكتم

لأاو هعرسلبا مكحتلا هاتج

آ تل ايل ج لأا بن ص ط ا بيج ماظنلا اذهو .لمتلمحا ماد ن

ىلع لمعي دعاسم ماظن ىلع يوتيح

إ ببستت دق تيلا تابقعلاو تابكرلما فاشتك ثراوكلا ثودحو مداصتلبا

نم هرداصلا تاراشلاا نوكت نأ بيج اضيا،

هدولجا لياع فاشتكلاا وا دصرلا ماظن يهو

لىولاا هوطلخا حتل

ي .بقترلما رطلخا دي ثحبلا اذهو

أ يمدقت لىا ىعس اضي

لماكتم ةاكامح ماظن ,

لاكب مكحتلا متيس ثيح نم

هعرسلا هاتجلااو . قناخ قيرط نع هعرسلبا مكحتلا اهيقشب هعرسلا

كرلمحا هو ي وقولا تىح هلماكلا هلمرفلا قيرط نع هعرسلبا مكحتلاو ،هعرسلا ليلقتل هوطلخا ف

كلذ دعبو . ماتلا تم

رايتخا Fuzzy control يهو

أ صلاا بنتج هتلوامح دنع ناسنلاا يركفت ةسارد ىلع دمتعت هيجيتاترس ط

ايتخاو ماد ر

أ و لضف أ ماظنلا في اهقيبطتو هلمتلمحا تارارقلا بسن قيرطلا هطيرخ نم تامولعلما يقتست ثيح

هفاسلماو إ صلاا لى ط و ماد

هروالمجا قرطلا في رورلما ةكرح لأا هبكرملل هعرسلاو

هيفللخاو هيمام ءاشنإ تم ايرخأ .

قرطلا يكايح تاراسم عبرا نم قيرطلا

.هيربكلاو هعيرسلا أ

تم اضي ءاشنإ يكاتح هدقعم تاهويرانيس ثلاث تاقرطلا في ثدتح دق تيلا هبعصلا فورظلا ضعب

.

ينتبكرم يهو جفو طسولاا قيرطلا في نايرست

أ جف قيرطلا في هبكرم لطعت ، امهماما هكرحتم يرغ هبقع رهظت ه أ

ه ، هرايس و

. هيربك هعرسب فللخا نم همداق ت ةاكالمحا جمنارب نم تاهويرانيس ثلاثلا ةجيتنو

نأ رهظ Fuzzy Control رداق

. حاجنب مادطصلاا بنتج ىلع

(4)

iv

APPROVAL PAGE

I certify that I have supervised and read this study and that in my opinion, it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Master of Science in (Automotive Engineering.)

………..

Waleed Fekry Faris Supervisor

………..

Faried Hasbullah Co-Supervisor

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Master of Science (Automotive Engineering).

………..

Fadly Jashi Darsivan Internal Examiner

………..

Mohamed Elsayed Okasha Internal Examiner

This dissertation was submitted to the Department of Mechanical Engineering and is accepted as a fulfilment of the requirement for the degree of Master of Science (Automotive Engineering).

………..

AKM Mohiuddin

Head, Department of Mechanical Engineering

This dissertation was submitted to the Kulliyyah of Engineering and is accepted as a fulfilment of the requirement for the degree of Master of Science (Automotive Engineering).

………..

Erry Yallian Triblas Adesta Dean, Kulliyyah of Engineering

(5)

v

DECLARATION

I hereby declare that this dissertation is the result of my own investigations, except where otherwise stated. I also declare that it has not been previously or concurrently submitted as a whole for any other degrees at IIUM or other institutions.

ALMUTAIRI SALEH RASHED

Signature ... Date ...

(6)

vi

COPYRIGHT PAGE

INTERNATIONAL ISLAMIC UNIVERSITY MALAYSIA

DECLARATION OF COPYRIGHT AND AFFIRMATION OF FAIR USE OF UNPUBLISHED RESEARCH

FUZZY- BASED COLLISION AVOIDANCE SYSYTEM FOR AUTONOUMOUS DRIVING IN COMPLICATED SCENARIOS

I declare that the copyright holders of this dissertation are jointly owned by the student and IIUM.

No part of this unpublished research may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without prior written permission of the copyright holder except as provided below

1. Any material contained in or derived from this unpublished research may be used by others in their writing with due acknowledgement.

2. IIUM or its library will have the right to make and transmit copies (print or electronic) for institutional and academic purposes.

3. The IIUM library will have the right to make, store in a retrieved system and supply copies of this unpublished research if requested by other universities and research libraries.

By signing this form, I acknowledged that I have read and understand the IIUM Intellectual Property Right and Commercialization policy.

Affirmed by ALMUTAIRI SALEH RASHED

……..……….. ………..

Signature Date

(7)

vii

ACKNOWLEDGEMENTS

Firstly, it is my utmost pleasure to dedicate this work to my dear parents, my wife and High Education Ministry in Saudi Arabia, who granted me the gift of their unwavering belief in my ability to accomplish this goal: thank you for your support and patience.

I wish to express my appreciation and thanks to those who provided their time, effort and support for this project. To the members of my dissertation committee, thank you for sticking with me.

Finally, a special thanks to Professor Waleed Faris for his continuous support, encouragement and leadership, and for that, I will be forever grateful.

(8)

viii

LIST OF TABLES

Table 2.1 Automotive Radar types based on Distance Measurement Ability 17 Table 2.2 Different Sensors with their Advantages and Disadvantages 19

Table 2.3 Classification of Time to Collision (TTC) 24

Table 3.1 Optimal Fuzzy Rules for Lane Change 42

Table 3.2 Optimal Fuzzy Rules for Adaptive Cruise Control 44

(12)

xii

LIST OF FIGURES

Figure 1.1A research methodology flow diagram 6

Figure 2.1 Automotive Safety Tree Diagram (Robert Bosch GmbH., 2006) 9 Figure 2.2 A classification of Active Safety System based on objectives (Bengler et

al., 2014) 11

Figure 3.1 The Car Coordinate System 30

Figure 3.2 Detection Finder System 31

Figure 3.3 Fuzzy Control System 32

Figure 3.4 Fuzzy logic states overview map 34

Figure 3.5 MF functions for the first sate (Front obstacle distance) 35 Figure 3.6 MF for the second state (rear object velocity) 36

Figure 3.7 MF for the third state (Current Lane) 37

Figure 3.8 MF for the fourth state (X location) 37

Figure 3.9 MF functions for the fifth state (left obstacle distance) 38 Figure 3.10 MF for the sixth state (right obstacle distance) 39

Figure 3.11 Overview for output controls 39

Figure 3.12 MF for the first output (car speed) 40

Figure 3.13 MF for the second output (car direction) 41

Figure 3.14 MF for the third output (Car mode) 42

Figure 3.15 Car-mode reaction with Rear object velocity and front object distance

states 45

Figure 3.16 Output car-direction reaction with current-lane and front object distance

states 45

Figure 3.17 Output car-speed reaction with X-location and front object distance states.

46 Figure 3.18 Road map environment with different scenarios 47 Figure 3.19 Road map environment with different scenarios 2 47

(13)

xiii

Figure 3.20 First Scenario 48

Figure 3.21 Second Scenario 49

Figure 3.22 Third Scenario 49

Figure 4.1 Obstacle detection parameters for lift and front side 52

Figure 4.2 Simulation Result for Scenario 1 52

Figure 4.3 Red Car Velocity 53

Figure 4.4 Front Obstacle Velocity, Black Car 53

Figure 4.5 Obstacle no1- x location (broken location) 54

Figure 4.6 Car -x location 54

Figure 4.7 Car y-Location (Red Car) 55

Figure 4.8 Detections parameters for white car 56

Figure 4.9 Obstacle detection model for white car 57

Figure 4.11 Car x- location during lane changing 59

Figure 4.12 Car y- location during lane changing 59

Figure 4.13 Car Velocity During Lane Change 60

Figure 4.14 Car Steering Angle Velocity 60

Figure 4.15 Obstacle detection parameters for rear and front side 61

Figure 4.17 Car-y location (Black) 63

Figure 4.18 Car Steering Angle Velocity (Black car) 63

Figure 4.19 Rear Obstacle Velocity, Obstacle no 2 64

Figure 4.20 Car Velocity (black) 64

Figure 4.21 Front Obstacle Velocity, Obstacle no 1 65

(14)

xiv

LIST OF ABBREVIATIONS

DAS Driver Assistance System CMP Crash Prediction Modul LIDAR Light Detection and Ranging TTC Time to Collision

TH Time Headway

CAS Collision Avoidance System PMD Photonic Mixer Device ABS Antilock Braking System TSC Traction Control System ESC Electronic Stability Control

NHTSA National Highway Traffic Safety Administration UNECE United Nations Economic Commission of Europe ATC Air Traffic Control

TACAS Traffic Alert and Collision Avoidance System ARPA Automatic Radar Plotting Aids

IMO International Maritime organization AIS Automatic Identification System UAV Unmanned Autonomous Vehicle USV Unmanned Surface vessels UGV Unmanned Ground Vehicle ACC Adaptive Cruise Control IR Infrared Ray

ICC Intelligent Cruise Control

(15)

xv GPS Global Positioning System

PID Proportional integral Derivative Controller NMPC Nonlinear Model Predictive Controller MPC Model Predictive Controller

MF Membership Function RFD Range Finder Distance

(16)

xvi

LIST OF SYMBOLS

∅ Steering Angle 𝜃 Rear Axle Angle

𝑥₁ x-Coordinate of the center of the rear axle for the car 𝑥₂ y-Coordinate of the center of the rear axel for the car 𝑥₃ The heading of the car

𝑥₄ Steering Angle of the car 𝑙 The length between axels

𝑢₁ Car velocity

𝑢₂ Steering Angle velocity 𝑑₁ Front obstacle distance 𝑑₅ Rear object velocity 𝓁 Current lane

𝑑₃ Left obstacle distance 𝑑₇ Right obstacle distance 𝑢₁ Car Speed

𝑠 Car mode

𝑥_2′ Car direction from control output 𝐺𝐻𝑍 Gigahertz

m meter

(17)

1

CHAPTER ONE INTRODUCTION

1.1 OVERVIEW

A collision free and safe journey is a demand for every single individual. Collision avoidance system has been practised very commonly for a few decades specially in naval and aerospace fields with the help of radar guidance system. Nowadays, researchers are showing a great interest on radar guided collision avoidance in automobile applications as well. It is highly challenging to design such a collision avoidance system that may have the capability on maintaining proper balance on all sorts of collision avoidance at the same time.

In this work, has been designed a suitable control algorithm for collision avoidance that can ensure proper decision-making capability. The control algorithm is designed under the consideration of uncertainties in proper decision making with handling multiple obstacle at the same time. In order to do that, a fuzzy controller has been designed for collision avoidance control algorithm. Fuzzy controller takes a decision depending on system defined rules. The fuzzy controller can take decision with the help of some pre-defined functions which include filters to avoid noise and disturbances to the system and takes certain inputs and executes outputs according to the desired response. Thus, collision avoidance control algorithm executes the final action depending on the fuzzy controller.

A MATLAB and Simulink environment has been chosen to design the controller and perform some simulations under certain conditions in order to estimate the performance of the controller.

(18)

2

1.2 STATEMENT OF THE PROBLEM AND ITS SIGNIFICANCE

Since human behaviour is the main cause of an accident, and the most common and damaging accident is the face to face accident according to the literature review in next chapter. The problem statement is how to improve driving behaviour to avoid road accident in most common accident area which is the front side.

To improve driving behaviour to avoid the accident its must go beyond human action and produce an assistance system based on host vehicle movement to assist the driver in dangerous situation. However, this system must be efficient, does not need human interaction and completely automated.

1.3 RESEARCH OBJECTIVES

1. Develop a system for autonomous driving in complicated driving situations;

2. To Select a mathematical model for vehicle proximity;

4. To design control algorithm for proximity detection and vehicle braking;

1.4 AUTOMOTIVE SAFETY

Safety is nowadays a prime concern of vehicle design among the vehicle designers and researchers. Therefore, the development of safety system in automobile is necessarily required in order to minimize the accident and damages after accidents. Designers could show their successes by minimizing the injury and death rates in developed countries through the improvement in automobile industry and roadway design (Isermann, Mannale, & Schmitt, 2012). The development has taken place in active and passive safety systems on automobile industry. Active safety system describes the system that can assist the driver to avoid any possible accident while passive system indicates the

(19)

3

features of vehicle that can help to minimize the effects of an accident. (Tongeren, Gietelink, Schutter, & Verhaegen, 2007).

Some advanced Driver Assistance System (DAS) features like traction control system and break assistance, collision avoidance systems, warning systems and …..etc.

could able to show the success by decreasing the automobile accident rates in some countries (Mukhtar, Xia, & Tang, 2015). According to DAS, safety distance is defined as a minimum distance that is required to be maintained in order to avoid collision with the following vehicle’s rear end (Raza & Ioannou, 1996). Therefore, DAS introduces with Crash Prediction Module (CPM) or Accident Prediction System that is developed in order to ensure the safety of passengers by predicting any probable accident in advance with the help of considering the relative velocity of vehicles. Velocity and relative positioning of the vehicles around are considered as the main traffic parameters for this system (Sreevishakh & Dhanure, 2015). So, whenever the system can detect an object of interest, it tries to detect whether the vehicle is under safe zone or dangerous zone. Then if it finds that the vehicle is in dangerous zone and there is a possibility of an accident, it informs the drivers about it and in autonomous vehicle cases, it informs the controller to take the vehicle in safe zone (Adla, Al-Holou, Murad, & Bazzi, 2013).

As it is required to predict about probable accident by estimating distance between the vehicle and object continually, several methods and devices are adopted in automobile industry like IR detector, Millimeter Wave Radar Technology, LIDAR and Vision based techniques (Tian, Bi, Liu, & Du, 2013; Zhao, Mammeri, & Boukerche, 2015).

The accident prediction system is operated with the help of sensors and a well-designed controller that can process data from sensors and send it to processor (Sreevishakh &

Dhanure, 2015). In terms of their application criteria, several approaches are found as like collision warning and avoidance systems, autonomous intelligent cruise control

(20)

4

systems, artificial-intelligence based systems and expert knowledge-based systems while each system contains its own control method and structure. (E.Lissel (Volkswagen AG), P.Andreas (Volkswagen AG), R.Bergholz (Volkswagen AG), &

H.Weisser (Volkswagen AG), 1996; FARBER & Huang, 1995; Ioannou & Chien, 1993;

Jula, Kosmatopoulos, & Ioannou, 2000; Mimuro, Miichi, Maemura, & Hayafune, n.d.;

A. Niehaus & Stengel, 1991; Axel Niehaus, Niehaus, & Stengel, 1994; Olney, Wragg, Schumacher, & Landau, 1995; Valayden, Guinand, Bizouerne, Bouaziz, & Maurin, n.d.).

Collision warning and avoidance systems can be classified by Sensor-based collision avoidance systems and ITS-based collision avoidance systems on the basis of tracking system. Sensor-based collision avoidance systems comprise on board sensors with processors where the processors collect information from real time to warn drivers (Trivedi, Cheng, Childers, & Krotosky, 2004). On the other hand, ITS-based collision avoidance system is subject to wireless communication that may take place between vehicles (V2V) and vehicles to infrastructure (V2I) where infrastructure indicates the units on roadside (Adla et al., 2013).

Therefore, collision avoidance system is dependent on sensors or wireless communication system to operate the system. This system also continually tries to detect any nearest object that can be a cause of an accident similar to CPM. After the detection of an object with the help of sensing devices, it takes the help of a filter in order to make generated signal free from noise. Kalman filter can estimate the states and make the signal smoother in order that collision prediction module system can detect the positions accurately. Kalman filter basically estimate the current states with the help of recent observation and current estimation (Faragher, 2012; Hao, Jinfeng, Chao, & Yi, 2017).

(21)

5

The task of the artificial-intelligence system in collision avoidance is to plan an optimal path based on analyzing the traffic situation.

Autonomous intelligent cruise control systems are called also Adaptive cruise control which the system will be active all the time and will control the speed only, without steering to avoid any front collision based on distance between the vehicle and an obstacle and usually follow the theory of time to collision (TTC).

Another well-known method, Time to Collision (TTC) works based on the difference of the velocities and time interval in order to estimate any collision. TTC basically estimates a certain time afterwards a collision may occur if two considered vehicles maintain their same velocities (Lee, 1976). Once it is found if the vehicles are in danger zone by estimating TTC, the system takes action according to the TTC like breaking or minimizing the speed (Yoshida, Awano, Nagai, & Kamada, 2006).

This work tries to avert the front collision with the help of millimeter wave radar of host vehicle. In addition, TTC theory has been adopted so that it can be determined if the host vehicle is in danger zone. Furthermore, three different categories of distance have been introduced such as danger distance, warning distance and safe distance so that the diver may take action like reducing the speed or breaking according to his position. Automatic emergency breaking system has been discussed for any emergency case of the host vehicle as well.

1.5 RESEARCH METHODOLOGY

This section presents the plan for achieving the objectives discussed in earlier section.

A literature review is an important section that can give an idea about the previous works. In this section, the difficulties and the problems which were faced was analyzed and discussed. Consequently, the analysis on the previous work on prevention control

(22)

6

system in indoor and outdoor environment can be considered as the primary step to achieve the goals. Secondly, a mathematical model will be selected which will be suitable for proximity vehicle movement. In order to build a complete system based on pre-selected mathematical model from previous step. MATLAB and Simulink software have been used to implement and check the performance of the designed controller.

Finally, a report will be written down that may contain all the simulated and practical results with a comprehensive discussion on the basis of the results.

Meet NO Requirement Selecting Control System Selecting Mathematical Model

Literature Review Start

Simulated Results

Change Parameters Model

Final Results and Reports

End

Figure 1.1A research methodology flow diagram YES

(23)

7 1.6 RESEARCH SCOPE

This research aims to develop a system for a car has the capability to avoid the collision in highway roads. The system should be fully automatic, using speed and manoeuvre to prevent the collision, have adaptive cruise control theory for controlling the speed without needs of lane changing, have the ability to reduce the speed until zero, change the lane if needs, has a detection subsystem that has the ability to detect any obstacles in 360˚ degree around the car with safe distance 1 mater which is suitable for down scale car , detecting rear obstacle velocity and can avoid the collision in a variously complicated traffic situation. This thesis focused more on the cars that travelling on high ways. Moreover, it is focusing on avoiding the collision by the understanding of human thinking when they are facing the imminent accident and transfer it to a control system, the control system should have the ability to achieve that.

(24)

8

CHAPTER TWO LITERATURE REVIEW

2.1 INTRODUCTION

The futuristic technological advancements nowadays have become indispensable part of our daily life. The advancement incorporates automobile industry as well that alludes both of its users and nonusers to be familiar with all the sophisticated features.

Therefore, it is also a matter of deep concern to the researchers in order to ensure the safety of the users. As a consequence, the modern cars are equipped with some additional devices and features to prevent any car theft incident, to avoid car accident and save drivers from any serious injury even if a crash takes place. The safety features that have been adopted in automobile industry and inferred by the researchers can be categorized by two different groups (Cloutier, Linke, & Bitar, 2011; Jansson & Unitr.), 2005):

1. Vehicle Security

2. Automotive Road safety 2.1.1 Vehicle Security

Automotive theft is one of the most persisting problem in Automotive industry around the world. It is reported that every year in America, one car is stolen in every twenty- eight seconds. Phelan et al. in their patent work develops an anti-theft model that incorporates a microprocessor-based control module. It only can enable solenoids’

operation of the power circuits if a correct identification code can be entered. The innovated device has a control on engine and some other important systems in order to limit or disable the systems (Phelan, KNS John - US Patent 5, 481, & 1996, n.d.). Guo et al. introduced with a secutrity system that can deter a thief to steal any car. Its

FUZZY- BASED COLLISION AVOIDANCE SYSTEM FOR AUTONOUMOUS DRIVING IN COMPLICATED