Online smooth error correction method for the reference trajectory that leads from an arbitrary initial state to the end state in positioning control

In this paper, a smooth error correction method for the reference trajectory in positioning control is proposed. In general, the reference trajectory that leads from an initial position to a destination is defined in the positioning control. When a change in the destination is considered during control, an error correction trajectory is necessary to smoothly lead to the new destination. Therefore, the reference trajectory that leads from an arbitrary initial state to the end state is derived. To verify the effectiveness of the proposed method, it is applied to vehicle longitudinal positioning control for an autonomous parking system. In general, image recognition is applied for a parking aid system to recognize a parking slot and determine a destination. However, determined destination usually includes errors caused by the distortion of the optical system by the incline or wave of the road surface. Generally, position detection is by image recognition as a recognition target approaches an optical axis center, and as the vehicle approaches destination, accuracy of detected position is enhanced. Thus, as the vehicle approaches its destination, accuracy of the detected position tends to improve in an autonomous parking system. It is shown through simulation that the proposed error correction method has smooth characteristics. Applying the proposed method, the longitudinal positioning accuracy is improved without losing ride comfort.