基于自适应扰动观测器的自主船舶协同路径跟踪控制

为实现未知环境扰动下不确定欠驱动自主船舶的协同路径跟踪控制, 本文提出了一种基于自适应扰动观 测器的鲁棒控制算法. 该算法采用径向基函数神经网络(RBFNNs)逼近模型参数不确定, 并利用最小学习参数化 (MLP)技术对神经网络的权重及逼近误差进行压缩, 所设计观测器不需要环境扰动上界的精确信息. 进一步, 基于 代数图论对船间通信进行建模, 设计了一种分散式协同控制律, 有效地降低了通信负载. 凭借Lyapunov稳定性理论 证明了闭环系统内信号的有界性, 且能通过对设计参数的调节使跟踪误差的收敛界为任意小. 最后采用数值仿真 试验验证了所提出算法的有效性和优越性.

Adaptive disturbance observer based cooperative path-following control for autonomous surface vessels

This paper proposed an adaptive disturbance observer based robust control algorithm to address the cooperative path following control of underactuated autonomous vessels under unknown time-varying environmental disturbance. In the algorithm, the radial basis function neural networks (RBFNNs) are employed to approximate the model parameter uncertainty. Based on the minimal learning parameterization (MLP) technique, both the weight and the approximation error of the neural networks are compressed. The disturbance observer is constructed without the information of the upper bound of the external disturbance. Furthermore, a decentralized cooperative control algorithm is presented on the basis of the algebraic graph theory, which reduce communication load between the autonomous vessels effectively. All signals in the closed-loop system are proved bounded by Lyapunov theory, and the bound of the error signal could be small enough by tuning the design parameters appropriately. Finally, numerical simulation is conducted to demonstrate the effectiveness and superiority of the proposed algorithm.