Nouri Nowrouz Mohammad, Valadi Mehrdad
Department of Mechanical Engineering, Iran University of Science and Technology, Tehran, Iran.
ISA Trans. 2017 Sep;70:288-297. doi: 10.1016/j.isatra.2017.02.006. Epub 2017 Jun 2.
Input design has a dominant role in developing the dynamic model of autonomous underwater vehicles (AUVs) through system identification. Optimal input design is the process of generating informative inputs that can be used to generate the good quality dynamic model of AUVs. In a problem with optimal input design, the desired input signal depends on the unknown system which is intended to be identified. In this paper, the input design approach which is robust to uncertainties in model parameters is used. The Bayesian robust design strategy is applied to design input signals for dynamic modeling of AUVs. The employed approach can design multiple inputs and apply constraints on an AUV system's inputs and outputs. Particle swarm optimization (PSO) is employed to solve the constraint robust optimization problem. The presented algorithm is used for designing the input signals for an AUV, and the estimate obtained by robust input design is compared with that of the optimal input design. According to the results, proposed input design can satisfy both robustness of constraints and optimality.
输入设计在通过系统辨识开发自主水下航行器（AUV）的动态模型中起着主导作用。最优输入设计是生成信息性输入的过程，这些输入可用于生成高质量的AUV动态模型。在最优输入设计问题中，期望的输入信号取决于要辨识的未知系统。本文采用了对模型参数不确定性具有鲁棒性的输入设计方法。贝叶斯鲁棒设计策略被应用于为AUV的动态建模设计输入信号。所采用的方法可以设计多个输入，并对AUV系统的输入和输出施加约束。采用粒子群优化（PSO）来解决约束鲁棒优化问题。所提出的算法用于为AUV设计输入信号，并将鲁棒输入设计得到的估计与最优输入设计的估计进行比较。根据结果，所提出的输入设计既能满足约束的鲁棒性又能满足最优性。
