Single-Link Flexible Arm Control Using Neural Dynamic Optimization Method

• Main Authors: Chien-Yi Wu, 吳建億
• Other Authors: Hsu-Jeng Liu
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/99001070652302063764

碩士 === 國立屏東科技大學 === 機械工程系 === 93 === The main feature of NDO is that it enables neural networks to approximate the optimal feedback solution, reducing the complexities of computation and storage problems of the classical methods such as dynamic programming (DP). Following some background and motivation about applying NDO on Single-Link Flexible Arm control, both linear and nonlinear models of Single-Link Flexible Arm dynamic system are presented. Particularly, we employ a conjugate gradient steepest descent algorithm for the NDO method to approximate the optimal control solution. The simulation results demonstrate that the NDO method can control the Single-Link Flexible Arm more effectively, with less iterations in comparison with a Backpropagation Gradient Decent Algorithm.