Neural network control of non minimum phase systems based on a noncausal inverse
Register, Andrew H.
Book, Wayne John
Alford, Cecil Orie
MetadataShow full item record
A new approach for feedforward ANN control of nonminimum phase mechanical systems is proposed. A standard backpropagation-of-errors ANN is used to form an inverse model controller which is applied to simulated nonminimum phase systems. Learning in the new approach is based on the convolution between a noncausal impulse response and a desired tip trajectory. Selection of the proper input set, input scaling and the ANN structure are investigated. Once the input and structure are specified, the ANN is trained over a single trajectory. After training, the ANN is used to drive the system in an open-loop configuration. Plots of the system states resulting from the ideal excitation and from ANN excitation are compared. The results obtained by varying both the number of units and the input set are presented. The results demonstrate the effectiveness of the proposed ANN inverse model approach.