中国造船
中國造船
중국조선
SHIPBUILDING OF CHINA
2013年
3期
58-69
,共12页
谢文博%付明玉%张健%施小成
謝文博%付明玉%張健%施小成
사문박%부명옥%장건%시소성
逆最优控制%动力定位%循迹%反步法%自适应算法
逆最優控製%動力定位%循跡%反步法%自適應算法
역최우공제%동력정위%순적%반보법%자괄응산법
inverse optimal control%dynamic positioning%tracking%backstepping%adaptive algorithms
针对动力定位船舶的循迹任务,提出了一种基于逆最优思想的自适应反步控制器,并对误差系统的稳定性进行了分析。首先对无干扰的动力定位船舶线性化模型设计最优跟踪控制器,利用时变Riccati方程进行反步状态变换得到误差系统,再基于 Lyapunov 函数对定常干扰设计自适应律和使系统渐近稳定的控制律,最后给出此控制律作用下的最优性能指标。该方法不用论证带有H∞范数的广义Riccati代数方程可解性,避免了传统非线性 H∞鲁棒控制需要求解 Hanmilton-Jacobi-Issacs(HJI)方程的问题。动力定位船舶的循迹控制仿真研究验证了算法的有效性。
針對動力定位船舶的循跡任務,提齣瞭一種基于逆最優思想的自適應反步控製器,併對誤差繫統的穩定性進行瞭分析。首先對無榦擾的動力定位船舶線性化模型設計最優跟蹤控製器,利用時變Riccati方程進行反步狀態變換得到誤差繫統,再基于 Lyapunov 函數對定常榦擾設計自適應律和使繫統漸近穩定的控製律,最後給齣此控製律作用下的最優性能指標。該方法不用論證帶有H∞範數的廣義Riccati代數方程可解性,避免瞭傳統非線性 H∞魯棒控製需要求解 Hanmilton-Jacobi-Issacs(HJI)方程的問題。動力定位船舶的循跡控製倣真研究驗證瞭算法的有效性。
침대동력정위선박적순적임무,제출료일충기우역최우사상적자괄응반보공제기,병대오차계통적은정성진행료분석。수선대무간우적동력정위선박선성화모형설계최우근종공제기,이용시변Riccati방정진행반보상태변환득도오차계통,재기우 Lyapunov 함수대정상간우설계자괄응률화사계통점근은정적공제률,최후급출차공제률작용하적최우성능지표。해방법불용론증대유H∞범수적엄의Riccati대수방정가해성,피면료전통비선성 H∞로봉공제수요구해 Hanmilton-Jacobi-Issacs(HJI)방정적문제。동력정위선박적순적공제방진연구험증료산법적유효성。
For the tracking task of dynamic positioning vessel, an adaptive backstepping control method based on inverse optimal idea is proposed; stability of error system is analyzed. First an optimal path following controller is designed for the linearized system of dynamic positioning ship without disturbance. The error system is obtained through backstepping transformation procedure using time varying Riccati equation. Adaptive law for estimating steady disturbance and control law which can make the system stable asymptotically is designed based on Lyapunov function. Finally optimal performance index is given under the given control law. It is not necessary to prove the existence of solution for generalized algebraic Riccati Equation with H∞norm, while solving Hanmilton-Jacobi-Issacs equation in traditional nonlinear H∞controller design is also avoided. By means of tracking control simulation for dynamically positioned vessels the effectiveness of the algorithm is verified.