ONLINE NEURO-TRACKING OF NONSTATIONARY MANUFACTURING PROCESSES

Two-phase self-organizing neuro-modeling (SONM), the global SONM and l ocal SONM, is designed for tracking non-stationary manufacturing proce sses. A radial basis function (RBF) neural network is employed, and a self-tuning estimator is also developed for the determination of RBF n etwork parameters on-line. A pattern recognition approach is presented for identifying a correct RBF neural network, which is used for ident ifying current manufacturing processes. Experimental results showed th at the proposed approach is suitable for tracking non-stationary proce sses. Copyright (C) 1996 Elsevier Science Ltd