지능형 통합 섀시 제어를 위한 종&#8226

Abstract

The automotive industry is currently confronted with even more stringent demands for active safety systems because their capabilities such as preventing fatal accidents through active intervention are increasingly being recognized in the market as a necessity rather than a luxury. Among such AVC (Advanced Vehicle Control) Systems and IUCC (Intelligent Unified Chassis Control) Systems, which assists the driver to keep the vehicle on the intended path and thereby helps to prevent accidents through differential braking, appears to have the most prosperousness. In particular, Information about vehicle velocity and sideslip angle is one of the most important factors for active safety systems, for enhancing vehicle stability control systems and also for achieving the intelligent vehicle initiative concept in the near future. Even if vehicle velocity and sideslip angle can be directly measured by instrumental sensors such as optical sensors or Global Positioning System (GPS) sensors, it is not feasible to install these sensors in passenger vehicles, because of high cost and lack of reliability. Therefore, an appropriate indirect sensing method that can overcome the above limitations is necessary for monitoring the lateral motion. The knowledge of tire-road friction coefficient can be also extremely useful to many active vehicle safety control systems including traction control, yaw stability control and roll-over prevention control systems. In particular, reliable estimation of the individual friction coefficient at each of the wheels will enable both traction and stability control systems to provide optimum drive torque and/or brake inputs to the individual. This dissertation focuses on the monitoring model and the monitoring system design for Intelligent Unified Chassis Control Systems. First, three monitoring systems are developed to improve the robustness in estimating lateral velocity and sideslip angle. Considering the fact that roll motion is closely related to the lateral tire forces in cornering, the first monitoring system is designed based on 2 DOF bicycle model including the rolling effect. The Sliding Mode Observer with Unknown Inputs (SMOUI) is designed to estimate the lateral velocity. The sideslip angle is determined by combining two estimation algorithms; Recursive Least Square with the Disturbance Observer (RLSDO) on the dynamic model and pseudo integral on the kinematic equation. The other two monitoring systems are designed based on the 3 DOF vehicle model including the longitudinal dynamics. The Extended Kalman Filter (EKF) including Dugoff’s tire model is designed to estimate longitudinal and lateral velocities. The Scaled Kalman Filter with Model Error Compensator (SKFMEC) technique is adopted in the third monitoring system to improve the robustness in estimating lateral velocity and sideslip angle. Secondly, this dissertation explores the development of monitoring algorithms for reliable estimation of the friction coefficient at each wheel and the maximum friction coefficient at front and rear wheels of the vehicle, respectively. The effects of pitch motion on the traction and braking tire force are taken into account to improve the accuracy of the longitudinal gyro sensor measurement. The dual Extended Kalman Filter (dEKF) using the estimated normal force and the maximum force method with parameter identifier synthesis is proposed based on the types of sensors available. The proposed monitoring systems are first evaluated in simulations and then evaluated experimentally with test vehicles. The estimation results demonstrate the feasibility of the proposed approaches for real-world applications.