CD IB fob r s / s 6 / f O ON INTERACTIVE CONTROL FOR INTELLIGENT COLLISION EVASIVE EMERGENCY INTERVENTION IN SMART VEHICLES L I B R A R Y UWVERSiTY OF MORATUWA. SRI L A M L f MORATUWA A thesis submitted to the Department of Electrical Engineering, University of Moratuwa in partial fulfillment of the requirements for the Degree of Master of Science by SAMARANATH RAVIPRIYA RANATUNGA Supervised by: Dr Sisil Kumarawadu University of Moratuwa 93949 Department of Electrical Engineering University of Moratuwa Sri Lanka 6 2 . \ - 3 ° ' ~TH J a n u a r y ^ ' ~ 9 3 9 4 9 Contents Declaration v Abstract vi Acknowledgement viii List of Figures ix List of Tables xi 1. Introduction 1 1.1 Collision Avoidance of Vehicles 1 1.2 Interactive Control of Vehicular Systems 1 1.3 Intelligent Applications in Vehicles 3 1.4 Hierarchy of Control among Vehicles 3 1.5 Inter-Vehicle Communication (IVC) 4 1.6 Modern Sensor Technologies for Smart Vehicular Systems 6 1.7 Main Controller Studies 7 1.8 Changes in Modes of Control and Human Factors Considerations 7 1.9 Simulation and Results 7 1.10 Prototype Studies/Realization of Prototypes 8 2. ANFIS: Adaptive Neuro-Fuzzy Inference Systems 9 2.1 Neuro-Fuzzy Hybrid Systems 9 2.2 Fuzzy Logic 9 2.3 Artificial Neural Networks 10 2.4 ANFIS Networks and Control 11 2.4.1 Hybrid Learning Algorithm 14 2.4.1.1 Least Squares Method 14 2.4.1.2 Gradient Decent Method 15 2.5 Summary 15 3. Auxiliary Functions: Synthesis of Controller 16 3.1 Vehicular Controllers 16 3.2 Auxiliary Functions 17 ii 3.2.1 Collision Condition Function 17 3.2.2 Relative Distance Function 19 3.2.3 Safety Speed Limit Function 20 3.2.4 Master-Slave Function 21 3.2.5 Steering Direction Function 23 3.3 Controller Algorithm 23 3.4 Summary 24 4. Synthesis of Controller and Training 25 4.1 Braking Controller 25 4.1.1 Fuzzy Input Membership Functions (Before Training) 25 4.1.2 Fuzzy Output Membership Functions (Before Training) 27 4.1.3 Training of the Braking Controller 28 4.1.4 Fuzzy Input Membership Functions (After Training) 28 4.1.5 Fuzzy Output Membership Functions (After Training) 31 4.2 Steering Controller 33 4.2.1 Fuzzy Input Membership Functions (Before Training) 33 4.2.2 Fuzzy Output Membership Functions (Before Training) 35 4.2.3 Training of the Steering Controller 35 4.2.4 Fuzzy Input Membership Functions (After Training) 36 4.2.5 Fuzzy Output Membership Functions (After Training) 39 4.3 Summary 39 5. Simulation Study 40 5.1 Vehicle Dynamic Model 40 5.2 Simulation Setup 44 5.3 Matlab/Simulink System Blocks for Simulation Setup 45 5.4 Parameters for each Vehicle in Simulation 48 5.5 Simulation Results 48 5.5.1 Side-End Collisions 48 5.5.2 Rear-End Collisions 52 5.5.3 Head-On Collisions 55 5.6 Summary 58 6. Prototype Realization 59 6.1 Component Selection 59 6.1.1 Controller Board-Oopic R+ 59 iii 6.1.2 Servo Motors 60 6.1.3 Radio Frequency Modules (Transmitter/ Receiver) 61 6.1.4 Ultrasonic Sensors 61 6.1.5 Digital Compass 63 6.1.6 Optical Encoder Modules 64 6.2 Testing of Individual Components for Realization of Prototypes 64 6.2.1 RF Module Testing 64 6.2.2 Servo Motors Calibration and Testing 66 6.2.3 Ultrasonic Sensor Testing 67 6.2.4 Digital Compass Testing 67 6.2.5 Optical Encoder Testing 67 6.3 Development of an Algorithm for Collision Avoidance Studies in Prototypes 68 6.3.1 Algorithm for the Prototype 69 6.3.1.1 Peripheral Obstacle Avoidance (OA) Module 69 6.3.1.2 Collision Avoidance (CA) Module 70 7. Conclusion and Future Directions 71 7.1 Conclusion 71 7.2 Suggestions for Future Directions 72 References 73 Appendices 76 Appendix A: Overview of Simulation Sub-System Blocks 76 Appendix B: Coefficients of the Trained Takagi-Sugeno Output Membership Functions of the ANFIS Braking Controller 83 Appendix C: Coefficients of the Trained Takagi-Sugeno Output Membership Functions of the ANFIS Steering Controller 87 Appendix D: Testing Program for RF Modules (On PC) 91 Appendix E: Testing Program for RF Modules (On Oopic R+) 93 Appendix F: Testing Program for Servo Motors 96 Appendix G: Programs for Checking Addresses of Ultrasonic Sensors 97 Appendix H: Testing Programs for Digital Compass 99 Appendix I: Testing Program for Optical Encoders 100 iv Declaration The work submitted in this thesis is the result of my own investigation, except where otherwise stated. It has not already been accepted for any other degree, and is also not being concurrently submitted for any other degree. S R Ranatunga ) (Candidate) Date: l e \ [ b \ / l o t f I endorse the declaration by the candidate. Dr. Sisil Kumarawadu (Supervisor) V Abstract This research study focuses on finding a solution for collision avoidance of smart vehicular systems. The main paradigm that is used to establish the solution is the interactive control of vehicular systems for negotiating a collision scenario for taking evasive actions. In this study, an interactive controller proposed negotiates collision scenarios between two vehicular systems leading to cooperative maneuvers. Thus, the interactive control actions lead to some maneuvers mutually beneficial to both the vehicles. The objective of this study is to develop a fully operational intelligent interactive controller for the smart vehicles. An Inter-Vehicle Communication (IVC) system plays a pivotal role in exchanging the necessary information in between the vehicles. The IVC system is assumed to be with enough versatility for dealing with multiple collisions in the channel transmitting information. This study is focusing on the vehicles outside the usually considered platoon environment. It is considering for emergency intervention maneuvers for collision evasive solutions. The hierarchical differentiation in control of the participatory vehicles is done by using the Master-Slave concept. The master is given more power in comparison to the slave. But these states are moment-bound and are to change fast. There are two main controllers which have been developed for braking and steering. The two controllers are based on Adaptive Neuro-Fuzzy Inference System (ANFIS). The top tier of this controller includes all important auxiliary functional components for processing the primary sensory variables. The ANFIS controller has been offline trained in the Matlab-7 environment. A simulation study has been done for the controllers in the Matlab/Simulink environment for various categories of collisions between the two vehicles. Even though the above paradigm is discussed for two participatory vehicular sub-systems, it is emphasized, that the same approach can effectively be extended without any major conceptual breakthrough to any number of vehicles for reliable evasion of collisions. In similar way, multiple vehicles can be considered as a multiplication of the number of pairs of vehicles for applying the results of the above study. vi Two fully autonomous prototypes were realized with full capability for testing intelligent interactive collision avoidance trials. Here, all sensor types and equipment were tested for expected functionality to be used in the integrated environment. To this end, software were developed for testing each component in the provided platform. vii Acknowledgement I specially thank my supervisor, Dr. Sisil Kumarawadu for his unwavering guidance, support and advice for carrying out this research work successfully. I am also very appreciative for his extensive help in fulfillment of some publications related to this research work, in some prestigious international forums. I am indebted to my parents and my wife for constant support and encouragement for successfully carrying out this work. My gratitude is also due to Prof. H.Y.R. Perera, Head/Electrical Engineering, for offering me a Research Assistantship in the Department of Electrical Engineering, in support of my studies. My sincere thanks are also due to the chairman and the committee members of the SRC grant committee, University of Moratuwa, for the grant they extended, which existed as an extensive support in my research studies. I would like to take this opportunity to extend my thanks to Dr. Amith Munindradasa, Dr. Rohan Munasingha, Dr. Thrishantha Nanayakkara and Dr. Lanka Udawatta for being the members of the review committee for my research. If not for their guidance and advice this work wouldn't have been a success at the end. Dr. N Munasinghe and his staff, at the Engineering Post Graduate Unit, are also thanked for all assistance extended. I have been assisted extensively by Mr. Geeth Jayendra, in realizing the prototypes for the research work. My appreciations are also due to him. I would also like to thank Mr. Buddhika Jayasekera and Mr. Dharshana Prasad, who have been my colleagues at the Departmental Research Lab, for helping me in various ways for successfully carrying out this work. Finally, my thanks go to various other personnel without whose help this work wouldn't have been a success. Understandably, their individual names cannot be mentioned here due to being large in number. viii List of Figures Figure Page Fig. 1 ANFIS controller common structure for braking and steering controllers 11 Fig. 2 Control system block diagram 16 Fig. 3 Detection of collision condition 18 Fig. 4 State diagram for change of control 22 Fig. 5(a) Membership functions of CollisionCondition_braking controller 25 Fig. 5(b) Membership functions of RelativeDistance braking controller 26 Fig. 5(c) Membership functions of MSSwitch_braking controller 26 Fig. 5(d) Membership functions of SpeedLimit_braking controller 26 Fig. 6(a) Membership functions for CollisionCondition_ braking controller after training 28 Fig. 6(b) Membership functions for RelativeDistance_braking controller after training 28 Fig. 6(c) Membership functions for MSSwitch_braking controller after training.... 29 Fig. 6(d) Membership functions for SafetySpeedLimit_braking controller after training 29 Fig. 6(e) Output control surface for the braking controller 30 Fig. 6(f) Root mean squared error for training ANFIS braking controller 31 Fig. 6(g) Root mean squared error for checking ANFIS_braking controller 31 Fig. 7(a) Membership functions of CollisionCondition_steering controller 34 Fig. 7(b) Membership functions of RelativeDistance_steering controller 34 Fig.7(c) Membership functions of MSSwitch_steering controller 34 Fig. 7(d) Membership functions of SteeringDirection_steering controller 35 Fig. 8(a) Membership functions for CollisionCondition_steering controller after training 36 Fig. 8(b) Membership functions for RelativeDistance_steering controller after training 37 Fig. 8(c) Membership functions for MSSwitch_steering controller after training... 37 Figure Page Fig. 8(d) Membership functions for SteeringDirection_steering controller after training 37 Fig. 8(e) Output control surface for the steering controller 38 Fig. 8(f) Root mean squared error for training ANFIS_steering controller 38 Fig. 8(g) Root mean squared error for checking ANFIS_steering controller 39 Fig. 9(a) Reference frames and position vectors 41 Fig. 9(b) Vehicle model 42 Fig. 10(a) Main Simulink block model for the simulation 45 Fig. 10(b) Sub-system controller block for the vehicle model-1(2) 46 Fig. 10(c) Main vehicle model sub-system with world-coordinate frame transformations 46 Fig. 11(a) Trajectories of the vehicles in near side-end collision scenario 49 Fig. 11 (b) Change of relative distance between the two vehicles 49 Fig. 11(c) Steering command (vehicle-1) 49 Fig. 11(d) Steering command (vehicle-2) 50 Fig. 11(e) Deceleration profile (vehicle-1) 50 Fig. 11(f) Deceleration profile (vehicle-2) 51 Fig. 12(a) Trajectories of the vehicles in near rear-end collision scenario 52 Fig. 12(b) Change of relative distance between the two vehicles 52 Fig. 12(c) Steering command (vehicle-1) 53 Fig. 12(d) Steering command (vehicle-2) 53 Fig. 12(e) Deceleration profile (vehicle-1) 54 Fig. 12(f) Deceleration profile (vehicle-2) 54 Fig. 13(a) Trajectories of the vehicles in near head-on collision scenario 55 Fig. 13(b) Change of relative distance between the two vehicles 55 Fig. 13(c) Steering command (vehicle-1) 56 Fig. 13(d) Steering command (vehicle-2) 56 Fig. 13(e) Deceleration profile (vehicle-1) 57 Fig. 13(f) Deceleration profile (vehicle-2) 57 Fig. 14 Oopic R+ controller board 59 Fig. 15 Hitec HS-422 standard deluxe servo motors 60 Fig. 16 RF communication modules: receiver and transmitter 61 Fig. 17 Beam pattern of the SRF235 'pencil beam' ultrasonic sensor 62 Figure Page Fig 18 SRF235 'pencil beam' ultrasonic sensor 62 Fig 19 Digital compass and external pin connections 63 Fig. 20 The optical encoder and code wheel 64 Fig. 21 Circuit diagram for RF communication between the PC and onboard RF modules 65 Fig. 22 Display for checking existing baud rate of the receiver 65 Fig. 23 Display after adjusting the baud rate to 9600 bps 66 Fig. 24 Pull-up resistors on HEDS-9040 encoder module outputs 67 Fig. 25 Prototype platforms with assembled components 68 List of Tables Table Page Table 1 Typical parameters (nominal) for each vehicle system 48 xi