水面无人艇无风险操纵运动控制研究

Within the field of marine traffic engineering, ship safety and maneuvering control are interdependent critical issues while traditional methodologies cannot effectively realize integrations, and thereby dramatically hindering theoretical and practical investigations. This proposal presents the integrative study on safety and control of unmanned surface vehicle (USV) based on theoretical principles of quaternion ship domains and self-organizing fuzzy neural networks we proposed in the previous study. As a consequence, traditional methodologies with separated study on the USV safety and maneuvering control would be innovated to present the novel USV collision risk-free maneuvering motion control theory. The significant contributions are as follows. 1) The model of navigation safety dynamics (NSD) is proposed by modeling the dynamics of intelligent quaternion ship domains. 2) The USV kinetics and dynamics would be combined with the NSD model, and thereby contributing to the generalized ship motion model (GSMM) with NSD, which would establish the promising model architecture for the integrated study on collision avoidance decision-making and maneuvering control. 3) Based on the GSMM, real-time intelligent collision avoidance decision-making algorithms are developed to propose the collision risk-free dynamic trajectory planning (DTP) satisfying COLREG rules. 4) Learning schemes of fast self-organizing fuzzy neural networks are presented to realize the online approximation for uncertainties and unmodeled dynamics of the GSMM. Combining with intelligent adaptive control theory, we reasonably propose the adaptive trajectory-tracking control based on the online intelligent approximation. This project would eventually establish systematic theories for the USV collision risk-free maneuvering motion modeling and control, and thereby contributing to substantial principles for marine intelligent traffic engineering.

船舶安全与操纵控制本是水上交通工程中两个共生的核心问题，而传统方法论并未将其有效融合，极大制约了该领域的理论研究和工程应用。基于四元船舶领域和自组织模糊神经网络等前期研究基础，本项目面向水面无人艇提出船舶安全与操纵控制一体化研究思路，开创水面无人艇无风险操纵运动控制研究。通过智能四元船舶领域动态建模研究，提出航行安全度动态模型；结合无人艇动力学和动态特性，提出具有航行安全度动态的广义船舶运动模型，建立避碰决策与操纵运动一体化研究的模型体系；基于广义船舶运动模型，研究实时智能避碰决策算法，提出具有无碰撞风险且满足海上避碰规则的动态航迹规划理论；基于快速自组织模糊神经网络学习算法，在线逼近广义船舶运动模型的不确定性和未建模动态，结合智能自适应控制理论研究，提出基于在线智能逼近的自适应航迹跟踪控制理论。本项目旨在建立水面无人艇无风险操纵运动建模与控制研究的系统化理论，为海上智能交通提供科学依据。

充分考虑水上交通和自然环境干扰等因素，结合船舶动力学和动态特性，建立了四元船舶领域动态模型；进而，针对动态变化的未知海况和交通状况，将前述模型推广至智能四元船舶领域动态模型，从而建立了本船相对于目标船或障碍物的船舶航行安全度动态模型。考虑时变交通状况和未知海况，以传统的船舶运动模型和航行安全度动态模型为子系统，将两者级联为具有航行安全度动态的广义船舶运动模型。采用所建立的具有航行安全度动态的广义船舶运动模型，研究了基于实时智能避碰算法的动态航迹规划理论。基于该实时智能避碰决策系统和广义船舶运动模型，提出了无碰撞风险最优航迹规划理论方法。结合在线模糊神经网络学习算法研究，开展了水面无人艇自适应航迹跟踪控制研究。通过提出快速在线自组织模糊神经网络学习算法，研究了广义船舶运动模型的在线辨识方法；基于该在线逼近模型，结合所得实时动态航迹规划算法和智能自适应控制理论，提出了水面无人艇智能自适应无风险航迹跟踪控制理论方法。