College of Navigation, Dalian Maritime University, Dalian 116026, China.
College of Navigation, Dalian Maritime University, Dalian 116026, China.
ISA Trans. 2023 Jul;138:49-62. doi: 10.1016/j.isatra.2023.03.011. Epub 2023 Mar 17.
Due to the harsh marine environment, the communication cost of multi-ship formation is expensive, but it is often ignored in the existing research. On this basis, this paper proposes a novel distributed anti-windup neural network (NN)-sliding mode formation controller of multi-ships with minimum cost. Firstly, distributed control is applied to devise the formation controller of multi-ships because it is a promising solution for the problem of single point failure. Secondly, the Dijkstra algorithm is introduced to optimize the communication topology, and then an optimized communication topology with minimum cost is used in the distributed formation controller design. Thirdly, to alleviate the influence of input saturation, an anti-windup mechanism is devised by combining an auxiliary design system with sliding mode control and radial basis function neural network method; and then a novel distributed anti-windup neural network-sliding mode formation controller of multi-ships is obtained, which can also handle the problem of nonlinearity, model uncertainty, and time-varying disturbances of ship motion. On the strength of Lyapunov theory, the closed-loop signals are proved to be stable. Multiple comparative simulations are carried out to validate the effectiveness and advantage of the proposed distributed formation controller.
由于海洋环境恶劣，多船编队的通信成本很高，但这在现有研究中往往被忽视。在此基础上，本文提出了一种新的基于最小代价的多船分布式抗饱和神经网络（NN）滑模编队控制器。首先，由于分布式控制是解决单点故障问题的一种很有前途的解决方案，因此本文将其应用于多船编队控制器的设计中。其次，引入了 Dijkstra 算法来优化通信拓扑，然后在分布式编队控制器设计中使用具有最小代价的优化通信拓扑。第三，为了减轻输入饱和的影响，本文通过将辅助设计系统与滑模控制和径向基函数神经网络方法相结合，设计了一种抗饱和机制；然后得到一种新的多船分布式抗饱和神经网络-滑模编队控制器，该控制器还可以处理船舶运动的非线性、模型不确定性和时变干扰问题。基于 Lyapunov 理论，证明了闭环信号的稳定性。进行了多次对比仿真，验证了所提出的分布式编队控制器的有效性和优势。
