SYSTEM IDENTIFICATION OF DYNAMICS OF UNDERWATER VEHICLE USING ARTIFICIAL NEURAL NETWORKS

海洋结构物、船舶和载人潜水器等水下机器人的动力学是复杂的、高度非线性的，不能用传统的动力学理论来考虑，特别是当它们在绝对低速运行时。此外，在操作过程中可能会改变动力学。为了处理这种复杂的时变动态，即使输入和输出是多个的，神经网络I/O系统也具有很好的学习能力。控制器和辨识模型由人工神经网络组成，控制器根据辨识模型的I/O关系进行自适应修正，提出了一种前馈神经网络结构及其学习过程来模拟被控对象的动态行为。该网络包括两种循环连接，即从输出层到输入层和从隐层到输入层。第一个连接使网络能够从自己的输出中获得输入状态变量，第二个连接是保持过去数据的影响。本文对学习过程进行了改进，使网络具有模拟包括高阶有限差分在内的动态行为的能力。将所提出的神经网络应用于水下机器人自适应控制系统--自组织神经网络控制系统中。SONCS中的神经网络控制器可以利用神经网络的仿真能力进行快速自适应。通过对多功能机器人Twin-Burger的航向保持控制的应用，成功地证明了该网络的有效性。

Dynamics of offshore structures, ships, and underwater vehicles such as manned submersibles are complicated and highly nonlinear to be considered with conventional dynamic theories, especially when they are operated in definitely slow speed. Moreover, the dynamics may be changed during operation. In order to deal with such a complex and time varying dynamics, the neural network I/O system is advantageous taking advantage of learning ability even if the input and the output are multiple. In this research, the controller and the identification model consist of the artificial neural network, and the controller is modified adaptively based on the I/O relation of the identification model.This year, a structure of feed forward neural network and its learning process were proposed to simulate the dynamic behavior of the controlled object. The network includes two kinds of recurrent connections, i.e., from the output layr to the input layr and from the hidden layr to the input layr. The first connection enables the network to obtain the input state variables from its own outputs and the second one is to keep the influence of the past data in itself. In this paper, the learning process is improved to equip the network with the capability of emulating the dynamic behavior including higher-order finite differences. The proposed network is adopted to the neural-network-based control system called "SONCS : Self-Organizing Neural-net Controller System" , which has been developed as an adaptive control system for Underwater Robots. The neural network controller in SONCS can be quickly adapted taking advantage of the network's simulating ability. The efficiency of the network is successfully demonstrated through the application to heading keeping control of a versatile robot called "Twin-Burger".