【作者】 吴恭兴；

【导师】 徐玉如；

【作者基本信息】 哈尔滨工程大学 ， 船舶与海洋结构物设计制造， 2011， 博士

【摘要】 水面无人艇是一种能够在实际海洋环境下安全自主航行,并完成各种任务的海上智能运动平台。与有人驾驶的舰艇相比,无人艇具有许多突出的战术技术特点,它体积小隐蔽性好、航速高、机动灵活、无人员伤亡危险等,在未来非对称立体战争中,无人艇将完成诸如扫雷、电子信息战、情报侦察与监视(ISR)、反恐、精确打击和武力保护等特殊任务；在恶劣海况(如海上大风、巨浪、热带风暴、海雾等)的探测研究和预警预报方面,无人艇将发挥其重要作用,因而越来越多地受到了世界各国的重视。在高度动态和不可预测的海洋环境里,为了达到高度自治,无人艇需要灵活可靠的操纵性能、精确快速的控制能力来确保其他船只与自身的安全；同时,从加强无人艇的自适应性与智能性,改善无人艇工作性能的观点来看,其控制系统还应该具备良好的自适应、自学能力功能,从而需要引入人工智能来设计无人艇的智能控制系统。本论文以“翔龙号”无人艇为研究对象,首先通过分析高速滑行艇的操纵运动机理,并结合实艇试验数据辨识的方法,建立了无人艇五自由度的操纵性数学模型；其次,利用最小二乘支持向量机的回归原理,根据结构风险最小化准则的学习原理,对小样本的无人艇Z型试验数据引入线性核函数,通过分步辨识的方法,辨识了无人艇艏向响应方程中的一、二阶线性与非线性参数,并验证了二阶非线性响应方程具有很好泛化性能；然后,在该无人艇操纵运动模型的基础上,把S面控制算法和模型参考自适应规律的设计原理相结合,根据李雅普诺夫稳定性理论设计了基于参考模型的S面自适应艏向控制器,并验证了该控制算法具有较好的抗模型参数变化和外界干扰的能力；之后,把仿人智能控制的在线特征辨识、启发式直觉推理逻辑和多模态控制的优点与人类小脑的协调运动机理相结合,提出了基于小脑模型的仿人智能协调控制策略,设计了基于发动机、喷水推进器喷嘴和倒车斗协调运动的无人艇基础运动控制策略,并根据人体运动控制系统结构的特点,提出一种仿人智能的混合式分层递阶结构,并利用智能图式的学习进化原理,在实践中不断完善和改进,则形成一种混合仿人智能图式的体系结构；最后,通过无人艇的定艏向、定航线等海上试验验证了该仿人智能协调控制方法和基于混合仿人智能图式体系结构的无人艇自主驾控系统的可靠性和工程实用性。更多还原

【Abstract】 Unmanned Surface Vehicle (USV) is one of intelligent motion platforms, which can navigate safely in the real marine environment and complete various tasks. Compared to the manned naval ships, the USV has a lot of outstanding tactical and technical characteristics, for instance, small volume, stealthy, high-speed, mobile and flexible, no dangerous about casualties and so on. The USV will complete special missions in the non-symmetric three-dimensional warfare, such as mine sweeping, electronic warfare, ISR (Intelligence, Surveillance and Reconnaissance), anti-terror, precision strike and force protection. And the USV will be useful in bad sea condition investigation and prediction, such as, gale at sea, billow, tropic storm, sea fog and so on. Therefore USV will be highlighted for special attention more and more around the world in the future.In the high-dynamic and uncertain ocean environment, in order to obtain highly autonomy, the USV must be endued with agile maneuver performance and precise control performance to insure the security for other ships and self. At the viewpoint of enhance the USV adaptability, intelligence, and functionality, the artificial intelligence should be introduced to designing the USV intelligent control system.In this thesis, the research object is "XIANG-LONG" USV. Firstly, the motion mechanism of the high-speed planing craft was analyzed, combine the system identification method, the five freedoms manoeuvring motion model was established. Secondly, according to the learning theory of structural risk minimize rule, the Least Squares Support Vector Machines (LS-SVM) based on linear kernel is adopted in the regression of the small stylebook zigzag motion data of the USV trial. Using the step by step identification method, the parameters were identified of the USV response models including the first-order linear, the first-order nonlinear, the second-order linear and the second-order nonlinear models. And the second-order nonlinear response model possessing well generalization was validated. Thirdly, on the base of the manoeuvring motion model of the USV, the S pane control arithmetic was combined with design theory of model reference adaptive rule; According to the Liapunov theorem of stability, the S pane adaptive controller was designed for the USV based on the model reference. And this control arithmetic was validated to be anti-parameter-changing and anti-jamming. And then the excellences of the human-simulated intelligent control, such as:characteristic identification on line, logic consequence of heuristic instinct and multi-mode control, were combined with the harmony motion mechanism of the human cerebella; The human-simulated intelligent harmony control strategy was proposed based on the cerebella model; And the USV motion control strategy was designed based on the harmony action of engine, the nozzle and deflector of the water-jet-propulsion. Fourthly, according to the structure characteristic of the human motion control system, a human-simulated intelligent hybrid hierarchical structure was proposed. Using the learning evolution theory of the human-simulated intelligent schema, this structure would be perfected and improved, then a system structure, named hybrid human simulated intelligent schema, would be formed. Finally, diversified trials were done at sea for the USV, such as, head keeping, and course tracking. The reliability and engineering practicability of the designed control method and designed system structure were validate by the USV trials results.更多还原