二輪車の動特性を考慮した車線追従制御に関する研究

抄録

This study employs Linear Quadratic (LQ) control as model-based control theory for a lane tracking controller of a two-wheeled vehicle. By examining simple computer simulation, lane tracking controller by using the LQ control shows a good lane tracking performance. The LQ controller makes it possible to realize stable lane tracking control by evaluating only directional (lateral motion) stability, while conventional model-free control theory for lane tracking control such as Proportional-Derivative (PD) control has to consider not only directional stability, but also trim (rolling motion) stability. In addition, evaluating the lateral displacement at the virtual point ahead of the vehicle, which is defined by synthesizing the lateral displacement at the center of gravity and the yawing angle, copes with both lane tracking performance and convergence. Adding a Kalman filter as a state estimator to the above-mentioned LQ controller, i.e. an LQG controller, the lane tracking performance is also realized using estimated values under the steering torque disturbance.