基于NSGA-II算法的水下机器人水平面运动参数辨识

摘要

水下机器人水动力系数的获得是分析其运动及设计控制器的基础。由于水下机器人水平面具有运动强耦合、非线性的特点，采用简单遗传算法来辨识水动力系数时无法使得多个目标函数同时达到最优，所以采用带精英策略非支配排序遗传算法(NSGA-II)来对水平面运动模型进行辨识。利用辨识得到的系数反求水下机器人水平面速度，仿真速度与实测速度曲线吻合较好，表明NSGA-II辨识得到的水动力系数是有效的。%Hydrodynamic coefficients are the basis of motion analysis and controller design for the underwater vehicle. The horizontal motion of the underwater vehicle is highly coupled and nonlinear, so it is difficult to achieve the optimal results for all the objective functions at the same time when using simple genetic algorithm to identify its coefficients. As an alternative, the elitist non-dominated sorting genetic algorithm (NSGA-II) is chosen to identify the hydrodynamic coefficients of the motion in the horizontal plane. The coefficients identified by NSGA-II are then used to calculate the velocities of the underwater vehicle and they are approach to the velocities measured by sensors on the underwater vehicle. It is proved that the method used is effective.