A Neural System for an Automatic Weighing

摘要

Multilayered perceptron neural networks are employed to model a weighing system to estimate the applied mass accurately while the system is still in the transient mode. This is achieved through the networks trained with the backpropagation and the extended delta-bar-delta algorithms to compare the performance of the networks. The results obtained from neural models show that the both neural models are found successful especially if the data sets are noisy.