【作者】 王显峰；

【导师】 赵希人；

【作者基本信息】 哈尔滨工程大学 ， 系统工程， 2009， 博士

【摘要】 船舶在海浪中航行时,在波浪扰动力和力矩的作用下必然会产生六自由度的摇摆运动。如果能够对船舶摇摆状态进行估计和预报,将对舰载机起降、船舶减摇、船舶对接等都具有重要意义。因此,船舶运动估计与预报一直是备受关注的课题。本论文结合国防基金预研项目“航母运动极短期预报及舰载机着舰引导姿态预报系统”和“水面舰船姿态预报控制及仿真研究”,基于鲁棒H∞滤波及奇异系统理论对船舶运动估计和预报问题进行研究。论文的研究内容如下：论文基于切片理论,运用流体力学知识对船舶运动模型中水动力系数进行了分析,并对水动力系数进行海浪谱加权处理,建立了基于海浪谱加权平均的船舶运动模型,为船舶运动估计与预报的研究提供了标称模型。在船舶运动水动力系数摄动情况下,通过对卡尔曼滤波误差的分析,论证了采用鲁棒H∞方法研究船舶运动估计和预报的必要性。论文提出了鲁棒H∞估计与预报方法,即综合考虑模型摄动和干扰统计特性未知的因素,针对船舶运动模型摄动设计稳定的滤波器与预报器,并使海浪干扰到船舶运动状态的估计误差和预报误差的衰减水平(H∞范数)小于给定值。主要包括以下三个方面的工作：1.对于船舶运动估计,利用线性矩阵不等式(LMI)方法,设计了基于二次稳定的船舶运动鲁棒H∞滤波器和基于参数依赖Lyapunov函数的船舶运动鲁棒H∞滤波器,并得到了滤波器的稳定条件及其参数矩阵的求解方法。二次稳定意义下的滤波器设计不需考虑系统参数摄动和干扰的具体性状,设计方法简单易行。但该方法要求对于所有允许的不确定参数存在统一的Lyapunov函数,导致具有较大的保守性。基于参数依赖的鲁棒H∞滤波器,是通过设计与系统的不确定性相关联的Lyapunov函数得到的,减少了设计的保守性,提高了船舶运动估计的精度。2.船舶运动系统存在奇异摄动现象,论文提出了基于奇异系统的鲁棒H∞估计器的设计方法,并给出了估计器存在的条件,即误差系统是正则、无脉冲和稳定的,利用LMI方法对参数矩阵进行求解。将该方法应用于船舶运动估计,对基于奇异系统理论的船舶运动估计问题进行了仿真研究,验证了该方法的有效性。3.对于船舶运动预报,提出了船舶运动鲁棒H∞预报器的设计方法,给出了预报器的稳定条件及其参数矩阵的求解过程。将鲁棒H∞预报问题转化为时滞系统的鲁棒H∞滤波问题,并根据时滞依赖和时滞独立的思想,分别设计了预报长度无关型和预报长度相关型鲁棒H∞预报器,利用LMI的线性目标函数最小化方法完成了H∞预报器的求解。该方法解决了模型摄动和海浪干扰统计特性未知时的船舶运动预报问题。更多还原

【Abstract】 Sailing in the ocean, ship inevitably produces 6-DOF swinging motion under wave disturbing force and disturbance torque. If the ship motion attitude can be estimated and predicted, it will bring great benefit to the take-off and landing of carrier aircraft, ship anti-rolling, ship mating and so on. Therefore, ship motion estimation and prediction have been attracting much attention. Supported by the National Defence Foundation "Extreme Short-time Prediction of Aircraft Carrier and Attitude Prediction System of Guiding Carrier Aircraft Landing" and "Research of Warship Attitude Prediction Control and Simulation", Problem of ship motion estimation and prediction is researched based on robust H∞method in this dissertation. The main contents of this dissertation are as follows:In this dissertation, the hydrodynamic force coefficients of the ship motion model is analyzed using the hydrodynamics knowledge based on the strip theory. The coefficients are weighted average by wave spectrum, and ship motion model based on weighted average by wave spectrum is established, which provides nominal model for the research of ship motion estimation and prediction. Under the circumstance with pertubing hydrodynamic force coefficients, estimation errors of Kalman filtering is analyzed, and the necessity of applying robust H∞method to research ship motion estimation and prediction is demonstrated.The methods of robust H∞estimation and prediction are proposed in the dissertation, that is, considering synthetically two aspects of perturbation in ship motion model and unknown uncertainty of disturbance input statistic characteristics, stable filter and predictor is designed to make the attenuation level of wave disturbance to estimation and prediction error (H∞norm) lower than setting value. The main contributions of this dissertation contain the following three aspects:Firstly, for the ship motion estimation, the ship motion robust H∞filter based on the quadratic stability and parameter-dependent Lyapunov function is designed via linear matrix inequality (LMI) method, and the filter stability condition and the solution method of the parameter matrices are obtained. The filter design based on the quadratic stability does not need to consider material characteristics of parameter perturbations and exogenous disturbance, such that the design method is easily implemented. However, the method requires that there is a unified Lyapunov function for all the permitted uncertain parameters, which leads to more conservativeness. The parameter-dependent robust H∞filter is abtained by designing Lyapunov function correlated with systmen uncertainty, which reduces the design conservativeness and enhances the precision of ship motion estimation.Secondly, there is singular perturbation phenomenon in ship motion system, thus the robust H∞estimator based on singular system theory and is developed. Existential condition of the estimator that the error system is regular, impulse-free and stabile is given, and its parameter matrices can be derived by resovling linear matrix inequalities. This estimator design method is simulated and applid to the ship motion estimation, simulation result verifies the validity of this method.Thirdly, ship motion robust H∞predictor is proposed to resolve the problem of ship mition prediction, and stability condition of the predictor and the solution method of parameter matriecs are given. It transforms the problem of robust H∞prediction into the problem of time-delay system robust H∞filtering. The distance-independent and distance-dependent robust H∞predictors are designed according to the delay-independent and delay-dependent ideas, and the solution of H∞, predictor is derived by minimizing linear object functions of LMI. The ship prediction problem with model perturbation and unknown disturbance statistic characteristics can be solved by the proposed method.更多还原