Title Page

Abstract

초록

Contents

Chapter 1. Introduction 11

Chapter 2. Model Construction for Sideslip Angle Estimation 15

2.1. Dynamic Model 17

2.2. Kinematic Model 20

Chapter 3. Finite Memory Sideslip Angle Estimation 23

3.1. Batch Form 23

3.2. Iterative Form 28

Chapter 4. Dynamic/Kinematic Model Fusion Using Neural Networks 31

4.1. Structure of Artificial Neural Network 31

4.2. Training Data 34

Chapter 5. Simulation and Experiments 35

5.1. Simulation 35

5.2. Experiments 39

Chapter 6. Conclusions 43

Bibliography 44

Table 5.1. Vehicle parameters used for simulation 36

Table 5.2. RTAMSE in the simulation 38

Table 5.3. RTAMSE in the experiment. 42

Figure 2.1. Lateral dynamics of bicycle model. 16

Figure 4.1. Structure of neural network learning systems for: (a) ANN1 and (b) ANN2. 33

Figure 4.2. Structure of dynamics/kinematics fusion algorithm using ANN 34

Figure 5.1. Change in vehicle velocity in the simulation. 37

Figure 5.2. Change in sideslip angle in the simulation. 37

Figure 5.3. Sideslip angle estimation error in the simulation 38

Figure 5.4. Korea International Circuit, where the experiment with a real vehicle was conducted. 40

Figure 5.5. Change in vehicle velocity in the experiment. 41

Figure 5.6. Change in sideslip angle in the experiment. 41

Figure 5.7. Sideslip angle estimation error in the experiment. 42