Neural Networks Modeling of Autonomous Underwater Vehicle

摘要：

This paper describes two different neural networksmodels for autonomous underwater vehicles (AUVs). The onlinemultilayer perceptron neural networks (OMLPNN) have beendesigned to perform modeling of AUVs of which the dynamicsare highly nonlinear and time varying. The online recurrentmultilayer perceptron neural networks (ORMLPNN) have beenadditionally designed to generate a memory to pervious statesand increase the performance of the modeling. The designedOMLPNN and ORMLPNN with the use of backpropagationlearning algorithm has advantages and robustness to model thehighly nonlinear functions. The proposed neural networksarchitectures have been designed to model the test bed for AUVnamed NPS AUV. Simulation results show effectiveness of theOMLPNN and ORMLPNN to deal with modeling of AUVs as ithas good capability to incorporate the dynamics of the system.

作者: R. Amin A. A. Khayyat K. G. Osgouie

会议类型: 国际会议

会议名称: 2010 IEEE/ASME International Conference on Mechatronic and Embedded System and Applications(2010 IEEE 机电一体化和嵌入式系统与应用国际会议)

会议地点: 青岛

会议语种:英文

页码: 14-19

在线出版日期: 2010-07-15（万方平台首次上网日期，不代表论文的发表时间）