抄録
We propose a method for vehicle path control that considers sideslip through the combination of a time-state control form and a path control method with nearest point search and we demonstrate the method's effectiveness through experiments with an experimental small four-wheeled vehicle. In the proposed method, combining the time-state control form and path control method with nearest point search for a geometric model that constrains the position of the vehicle in consideration of sideslip allows arbitrary path following through the translational velocity and yaw rate determined from input. The method gives steering angle as input to the yaw rate control system with derived yaw rate taken as a desire value for an equivalent two-wheeled model that considers sideslip. Because cornering power and sideslip angle are difficult to determine in a yaw rate control system, these are estimated using adaptive identification and a state observer, and an adaptive control system is formed. The small four-wheeled vehicle performs path control by use a laser scanner to estimate its position and attitude. The effectiveness of a sideslip angle estimation method that uses motion capture is also demonstrated. We verify that estimates of sideslip angle are nearly identical to measured values, and experiments using a variety of path desired values demonstrate the effectiveness of the proposed path control method.
