CAJ | 학술논문

在外界干扰和参数不确定的情况下，设计一种基于Backstepping的自适应神经滑模控制器对欠驱动船舶神经滑模航迹进行控制。采用趋近律的滑模控制，抑制常规滑模的固定切换增益系数带来的抖振现象；利用神经网络辨识对象，减少趋近律滑模控制对对象模型参数的依赖。以某实习船为例进行仿真，其结果表明：在标称参数下，所设计的控制器能够有效跟踪设定的航迹并抑制常规滑模控制器的抖振现象；在存在外部环境扰动以及参数摄动的情况下，仍然能够实现较好的控制，表现出强鲁棒性。
재외계간우화삼수불학정적정황하，설계일충기우Backstepping적자괄응신경활모공제기대흠구동선박신경활모항적진행공제。채용추근률적활모공제，억제상규활모적고정절환증익계수대래적두진현상；이용신경망락변식대상，감소추근률활모공제대대상모형삼수적의뢰。이모실습선위례진행방진，기결과표명：재표칭삼수하，소설계적공제기능구유효근종설정적항적병억제상규활모공제기적두진현상；재존재외부배경우동이급삼수섭동적정황하，잉연능구실현교호적공제，표현출강로봉성。
As for uncertain external interference and parameters, a self-adaptive neural sliding mode controller based on backstep-ping is designed for underactuated ships powered by wind, wave and flow. The chattering problem brought on by fixed switching coeffi-cient of traditional sliding mode control ( SMC) is restrained by using approaching-law SMC. And the neural network is utilized to re-duce the dependence on plant model’ s parameters in approaching-law SMC. Results of a ship simulation indicate that the controller can effectively keep the given track and avoid the chattering phenomenon occurring in traditional sliding mode control and it can also control well under the external disturbance and parameter perturbation, which shows the controller strong robust.