基于模型预测控制的自主水下机器人回收控制算法

Docking Control Algorithm for AUV Based on Model Predictive Control

针对自主水下机器人在水面回收控制中遇到的系统强耦合、多约束问题,提出一种基于模型预测控制的对接回收控制算法。通过对自主水下机器人动力学方程进行解耦,给出了水平面运动的动力学模型;对运动过程中的可见性约束和系统的输入约束进行公式化表述,使用模型预测控制框架进行带约束的最优化问题滚动求解,给出最优的系统控制输入。使用REMUS水下机器人的参数进行了MATLAB仿真,并根据对回坞导引控制性能的影响给出了最佳的预测时域,研究结果表明,该方法能够保证回坞过程中的可见性约束,回坞过程收敛性强、精度高,能够快速高效地完成自主水下机器人的水面回收。

To solve the problem of strong coupling and multiple constrains of horizontal docking of autonomous underwater vehicle,the autonomous docking method based on model predictive control is proposed. By decoupling the dynamic equations,the horizontal dynamic model of autonomous underwater vehicle is proposed. The visibility constraints and system input constraints in the docking process are formulized. Using model predictive control scheme,the receding solving of the optimal control problem with multiple constraints are conducted,and thus,the optimal system control input is provided. The simulations on MATLAB are carried out with REMUS AUV parameters,and the best prediction domain step time is chosen according to the system performance. The results show that the method is of strong convergence and high precision,while the visibility constraints are always satisfied,such that the horizontal docking of autonomous underwater vehicle is achieved rapidly and effectively.