AUV Vertical Motion Control Based on Kalman Filtering

Abstract

AUV docking is a hot topic in underwater robot research. In order to fulfill the mission of docking, AUV needs to have more precise vertical navigation control ability, reduce the depth of the sensor there is a big noise data when calculating the error and AUV vertical depth when motion is not smooth. In this paper, the kalman filter is integrated into the motion control of vertical plane, and the double closed-loop PID cascade control system is designed and not based on the model. The whole control system is divided into two loops, the inner ring for the trim Angle PID controller, the output through the thrust allocation to calculate the required torque and torque, outer ring for the depth of the PID controller, the output for the input of pitch Angle. The kalman filter is integrated into the feedback loop of the depth data to improve the accuracy of the feedback data. The precision of vertical motion control is reflected by the stability of fixed depth navigation. Through the experiment on the lake, the depth mean square deviation of the vertical plane at the speed of 2kn is 0.24m2, the mean square deviation of the vertical Angle is 0.18 degree2 which proves the feasibility of this method.