充分发挥申请人在无人艇编队控制、多智能体系统协同控制、不确定性系统智能控制、船舶电气推进方面的研究优势，针对具有高度非线性、不确定、强耦合、欠驱动、多推进方式等综合复杂性的异构无人艇集群系统，从建模、规划、控制一体化的角度出发，将深度强化学习融入到无人艇集群运动规划和控制框架内，揭示无人艇与复杂海洋环境动态交互过程中的航行模式和操纵性规律，建立以数据驱动为核心的无人艇类脑决策与控制方法，使得无人艇具有自组织、自学习、自优化能力，通过自学习掌握复杂海况下的协同运动技能，实现无需模型先验知识的高技能运动控制，显著提高未知海洋环境下无人艇协同作业的安全性、适应性、高效性。研制无人艇集群机器学习控制系统，对所提规划和控制方法的有效性进行实验验证。通过深入的研究，从理论和实践两方面推动无人艇集群类脑决策与控制技术的发展，为无人艇海洋装备由远程操控级、协同交互级向自主学习级智能演进提供核心支撑。

The principal investigator of this project is experienced in formation control of unmanned surface vehicles (USVs), coordinated control of multi-agent systems, intelligent control of uncertain systems, and electrical propulsion of ships. Taking this advantage, this project focuses on a swarm of heterogeneous USVs subject to integrated complexity in terms of high nonlinearity, uncertainty, strong coupling, under-actuation, various propulsion ways, etc. By the integration of modeling, planning and control, and the introduction of deep reinforcement learning into the motion planning and control framework of USVs, this project will uncover the moving pattern and maneuvering principle during the dynamic interplay of USVs and complex sea environment, and build the data-driven brain-inspired decision and control method for USVs, enabling the capabilities of self- organization, self-learning, and self-optimization. Through learning, several coordinated motion skills will be mastered by the USVs, highly skilled motion control will be achieved without human knowledge of modeling, and the adaptation, safety, and efficiency of USVs under unknown sea environment will be enhanced. This project also aims to develop machine learning and control systems for the swarm of USVs, and verify the effectiveness of the planning and control method. Through deep research, the completion of this project will advance the frontiers on brain-inspired decision and control of a swarm of USVs, and provide the core technical support for intelligent evolution of USVs from remote control and coordinated interaction to autonomous learning.