Corpus: \bank\HKEC\engineering.txt at position 53406068
is studied by using a heater PT326. Finally, the paper is concluded in section 4.
Neural Control Algorithm
Two approaches are widely used for design of adaptive control: Direct and Indirect[3]. In this paper an indirect approach is implemented[4]. Two muiltilayer neural networks(MNNs) are required, one is the neural Plant Emulator(PE) and the other is the Neural Controller(NC). The emulator is off-line trained with the data set sufficiently rich to allow system identification. The Neural Controller has been proposed to learn the inverse dynamics of the plant from the observation of its input-output relationship, and then both NC and PE are on-line trained.
Training Algorithms
Use the BACK button to return to the previous page.
Chris Greaves, Web Concordancer
All rights reserved
