Cooperative-intelligent transportation systems (C-ITS) are a powerful solution to handle the problems of the transportation sector such as traffic incidents, traffic congestions, air pollution, and global warming. Ultimately, C-ITS have been developed for the safety and efficiencies of road and fuel. Using advanced vehicular communications such as dedicated short-range communication (DSRC) or cellular communication, vehicles and road infrastructures cooperate for the safety and the efficiencies, unlike the traditional intelligent transportation systems (ITS). Therefore, in C-ITS, it is important for the vehicles and the road infrastructures to retrieve necessary information in a timely manner. To achieve several goals of C-ITS (e.g., the safety and the efficiencies of road and fuel), we aim for service efficiency, fuel efficiency, and effective attack detection. To this end, we propose the following approaches. First, we introduce a data dissemination system in a bidirectional road scenario for efficient data services. Second, we propose an eco-driving guidance and eco-signal system to reduce fuel consumption and improve traffic flow at signalized intersections. Third, we investigate speed harmonization and merge control to manage the mixed traffic that consists of human-driven vehicles and connected automated vehicles (CAVs) at the bottleneck areas on highways. Finally, we propose a method to detect malicious information attacks in platoons. Through a field or realistic simulation test, we evaluate the performance of our approaches and demonstrate the trustable results. Our proposed data dissemination system contributed to improving the service efficiency by enhancing vehicle-to-vehicle (V2V) data sharing. We effectively reduced the fuel consumption at the signalized intersection, comparing to a scenario without the eco-guidance, improving the traffic flow using the eco-signal mechanism. In the merge areas on highways, CAVs effectively reduced the fuel consumption by controlling the arrival speed of mixed traffic using the speed harmonization; the merge control alleviated the congestion level by assigning priority to vehicles at the merge area. Using our proposed attack detection method, we can quickly detect various attacks regarding attack duration, falsification size, and falsified information. To sum up, our approaches enhance the safety and improve the efficiencies of fuel and data service effectively, in realizing the potential of safe and efficient C-ITS.
Table Of Contents
1 Introduction 1 1.1 Motivation 1 1.2 Objectives 3 1.3 Challenges 6 1.4 Approaches 8 1.5 Dissertation Contributions 10 1.5.1 RSU-assisted Adaptive Scheduling for Vehicle-to-Vehicle Data Sharing in Bidirectional Road Scenarios 11 1.5.2 Field Evaluation of Vehicle to Infrastructure Communication-Based Eco-Driving Guidance and Eco-Signal System 12 1.5.3 Speed harmonization and merge control using connected automated vehicles on a highway lane closure: A reinforcement learning approach 13 1.5.4 An Approach to Detecting Malicious Information Attacks for Platoon Safety 14 1.6 Dissertation Organization 14 2 RSU-assisted Adaptive Scheduling for Vehicle-to-Vehicle Data Sharing in Bidirectional Road Scenarios 16 2.1 Introduction 16 2.2 Related Work 20 2.3 System Architecture 23 2.4 Hybrid Centralized and Ad hoc data scheduling (HCA) problem 28 2.5 RSU Cooperation-based Adaptive Scheduling (RCAS) Algorithm 31 2.5.1 Centralized data scheduling 32 2.5.2 Dynamic clustering mechanism 38 2.5.3 Ad hoc data scheduling via V2V communication 43 2.6 Performance Evaluation 45 2.6.1 Setup 45 2.6.2 Simulation Results 48 2.7 Summary 53 3 Field Evaluation of Vehicle to Infrastructure Communication based Eco-driving Guidance and Eco-signal System 54 3.1 Introduction 54 3.2 Proposed System 57 3.2.1 System architecture 58 3.2.2 Eco-guidance and eco-signal mechanisms 59 3.3 Evaluations and Results 67 3.3.1 Test scenarios 68 3.3.2 Test results 71 3.4 Summary 78 4 Speed harmonization and merge control using connected automated vehicles on a highway lane closure: A reinforcement learning approach 81 4.1 Introduction 81 4.2 Related Works 86 4.3 Reinforcement Learning 89 4.3.1 Q-Learning 89 4.3.2 DQN 91 4.4 Methodology 92 4.4.1 Test Environment of Vehicles and Highway 93 4.4.2 DQN Framework 94 4.5 Performance Evaluation 106 4.5.1 Simulation Setup 106 4.5.2 State of the Practice Algorithms 108 4.5.3 Performance Evaluation 109 4.6 Summary 118 5 An Approach to Detecting Malicious Information Attacks for Platoon Safety 120 5.1 Introduction 120 5.2 Related Work 123 5.3 Platoon and Attack Models 127 5.3.1 Platoon model 128 5.3.2 Attack models 130 5.4 Methodology 133 5.4.1 Architecture of LMID 133 5.4.2 Training/Test data sets 135 5.5 Performance Evaluation 137 5.5.1 Setup 138 5.5.2 Evaluation results 141 5.6 Summary 149 6 Conclusions 151 References 164 Summary in Korean 180