張堂賢臺灣大學：土木工程學研究所王介Wang, ChiehChiehWang2007-11-252018-07-092007-11-252018-07-092007http://ntur.lib.ntu.edu.tw//handle/246246/50159本研究之目的為建立一個整合警告與控制之先進安全車輛(Advanced Safety Vehicle, ASV)車載系統架構。此架構包含危險駕駛警示系統(Inattentive Driving Warning System, IDWS)以及適應性巡航控制系統(Adaptive Cruise Control System, ACCS)。本架構主要利用暨有影像處理系統作為資料取得之基礎，並針對取得之資料加以運算、分析及處理，進一步發展危險駕駛警示與適應性巡航控制之警示與控制邏輯，並開發系統軟體實作。 危險駕駛係指駕駛人因精神不良、與乘客交談、使用手機、飲食、更換音樂CD、查看衛星導航地圖等，導致駕駛人沒有專心注意路況所產生之駕駛行為。本研究所發展之危險駕駛警示系統(IDWS)，主要包含車道偏離警示系統(Lane Departure Warning System, LDWS)與追撞警示系統(Rear-end Collision Warning System, RCWS)。車道偏離警示系統(LDWS)主要偵測行駛道路之車道線，並利用車輛與車道線之相對關係，取得車輛在車道當中的位置。應用 - 濾波器針對所取得之資料加以動態預測，並建立機率類神經網路(Probabilistic Neural Networks, PNN)，以判斷車輛之側向行駛軌跡為偏離車道(lane departure)或是變換車道(lane change)，若判斷結果為車輛即將偏離車道，即發布警告訊息。追撞警示系統(RCWS)則係藉由追蹤本車行駛車道前方之車輛(preceding vehicle)與本車(host vehicle)之相對關係，包括相對位置與相對速度，應用移動平均法與 - 濾波器針對所取得之資料加以動態預測，並配合駕駛反應時間，於車輛追撞前車前發出警告，提醒駕駛人及時作出反應，以期能降低危險事故發生之可能。 適應性巡航控制(ACC)發展自傳統定速巡航 (Conventional Cruise Control, CCC)，能自動控制車輛之縱向跟車距離，除了能達到定速巡航之目的，更能自動保持適當之跟車距離，以免與前車發生事故，或因啟動煞車而終止定速巡航系統之情況。本研究之適應性巡航控制系統(ACCS)，乃利用本車與前車之相對距離、速度、與加速度等資訊，動態模擬適應性巡航所應控制之油門方向角(throttle angle)以及煞車壓力(braking pressure)，並運用模糊理論作為控制之判斷邏輯，配合適應性控制理論，調整油門方向角(throttle angle)以及煞車壓力(braking pressure)之強度，以期發展適用之適應性巡航控制系統。 本研究使用車載螢幕、汽車音響與工業電腦作為系統硬體架構，利用Borland C++ Builder作為系統程式開發之軟體，並使用CCD（Charge Couple Device, 光電耦合元件）攝影機作為影像擷取設備，建立一個先進安全車輛(ASV)系統。本研究之預期結果乃是建立一個整合危險駕駛警示與適應性巡航控制之即時先進安全車輛(ASV)系統，能夠實際應用於道路駕駛情況，即時發出適當的警告與控制。The objective of this study is to develop an integrated ASV (advanced safety vehicle) on-board system that combines inattentive driving warning systems (IDWS) and adaptive cruise control system (ACCS). By applying image processing technology as the data acquiring medium, this study processes the acquired data and further calculates values of different parameters for the use of IDWS and ACCS. In chapter one we introduce the background, motivation, objectives, and structure of this study. In chapter two we address the details of research fundamentals such as data collection, camera calibration, and system configuration. Chapter three, we show the inattentive driving warning system, including lane departure warning system (LDWS) and rear-end collision warning system (RCWS) in detail. Moreover, for the LDWS and RCWS, we have off-line tests to see if these systems work. In chapter four, the concepts and theories of the adaptive cruise control system (ACCS) are defined. We have an off-line simulation for the ACCS as well. In chapter five, we have real-time experiments for the IDWS that really put our equipments on the vehicle and run the systems. The results of off-line and real-time experiments turned to be good that for off-line systems: LDWS had 100% accuracy; RCWS 96.43%. For real-time experiments, the LDWS had a precision of 89.29% and RCWS 80%. Both off-line and real-time experiments’ results were excellent, which is saying that the system developed in this study is a robust system.誌謝 i 摘要 iii Abstract v Table of Contents vii List of Figures ix List of Tables xi Chapter One Introduction 1 1.1 Background and Motivation 1 1.2 Objectives 3 1.3 System Framework 4 1.4 Literature Review 6 1.4.1 Real-time Lane Departure Warning System 6 1.4.2 Real-time Rear-end Collision Warning System 7 1.4.3 Adaptive Cruise Control Simulation System 8 1.5 Thesis Organization 9 Chapter Two Fundamentals 11 2.1 Introduction 11 2.2 System Configuration 11 2.2.1 Hardware 11 2.2.2 Software 15 2.2.3 Constraints 15 2.3 System Fundamentals 15 2.3.1 Vehicle Fixed Coordinate System 15 2.3.2 Image Coordinate System 16 2.3.3 Definition of Parameters 17 2.4 Camera Calibration 19 2.4.1 Introduction 19 2.4.2 Camera Calibration Experiment 19 2.4.3 Longitudinal Measurement Calibration 21 2.4.4 Lateral Measurement Calibration 23 2.5 Data Collection 27 2.5.1 Introduction 27 2.5.2 Lateral Measurements 29 2.5.3 Longitudinal Measurements 31 Chapter Three Real-time Inattentive Driving Warning System 35 3.1 Introduction 35 3.2 Lane Departure Warning System 36 3.2.1 Introduction 36 3.2.2 Detection of Lane Departure 36 3.2.3 Probabilistic Neural Networks 37 3.2.4 PNN Training 39 3.2.5 Scenario Classification Experiment 41 3.2.6 Results 53 3.2.7 Discussion 54 3.2.8 Summary 56 3.3 Rear-end Collision Warning System 59 3.3.1 Introduction 59 3.3.2 Detection of Near Rear-end Collision 59 3.3.3 Offline RCWS Tests 66 3.3.4 Discussion 70 3.3.5 Results 74 3.3.6 Summary 75 3.4 Summary 77 Chapter Four Adaptive Cruise Control System 79 4.1 Introduction 79 4.2 Fuzzy Controller 79 4.2.1 Fuzzy logic and control 81 4.2.2 Fuzzy Logic Switching Unit 83 4.2.3 Throttle Control Unit 89 4.2.4 Neutral State 91 4.2.5 Brake Control Unit 94 4.3 Adaptive Control System 96 4.4 ACCS Experiments 99 4.4.1 Introduction 99 4.4.2 Off-line Simulation 100 4.4.3 Simulation Results 101 4.4.4 Discussion 105 Chapter Five Experiments 107 5.1 Introduction 107 5.2 SYSTEM Overview 107 5.3 Experiment Design 110 5.3.1 Real-time LDWS Experiments Design 110 5.3.2 Real-time RCWS Experiments Design 111 5.4 Experimental Results and Discussions 112 5.4.1 Real-time LDWS Experimental Results and Discussions 112 5.4.2 Real-time RCWS Experimental Results and Discussions 115 5.5 Off-line and Real-time IDWS Tests Comparison 118 5.6 Summary 120 Chapter Six Conclusions and Recommendations 121 6.1 Conclusions 121 6.1.1 Conclusion for IDWS 121 6.1.2 Conclusion for ACCS 121 6.1.3 Summary 122 6.2 Recommendations 122 6.2.1 Recommendations for IDWS 122 6.2.2 Recommendations for ACCS 122 6.2.3 Summary 123 Reference 125 Appendix A 133 Appendix B 135en-US先進安全車輛車道偏離追撞適應性控制類神經網路模糊邏輯ASVlane departurerear-end collisionadaptive controlneural networksfuzzy logicthesis