| 要約: | Abstract Safe driving requires a comprehension of critical instances and dangers, achievable only through accurate estimation of dynamic vehicle behavior and road characteristics. This study proposes a Takagi-Sugeno (T-S) fuzzy functional observer capable of real-time estimation of unmeasured states (side-slip angle, yaw rate, angular displacement) and unknown inputs such as road curvature, using Lyapunov-Krasovskii stability theory and Linear Matrix Inequalities (LMIs) for parameter design. The objective is to provide a computationally efficient and robust solution that overcomes the restrictive assumptions of Proportional-Multiple Integral Observers (PMIO) and the complexity of Fuzzy Unknown Input Observers (FUIO) while ensuring real-time feasibility for embedded automotive systems. Comparative simulations and Processor-in-the-Loop (PIL) validation demonstrated that the proposed observer achieved superior accuracy, faster convergence, and lower computational cost than Full-Order Observer (FO), PMIO, and FUIO, confirming its novelty and practical potential for integration into advanced driver assistance systems to enhance safety and reduce accident risks.
|
|---|
