Synchronous motor drive modeling using entropy-based process reconstruction

This paper discusses the variables needed to represent the dynamics of a synchronous motor drive system. This is an important issue when automatic input/output modeling, based on learning through examples algorithms, such as neural networks, is considered. The design of a neural network requires, amongst other things, the proper choice of input variables, avoiding over fitting and an unnecessarily complex input vector. This may be achieved by trying to reduce the arbitrariness in the choice of endogenous variables. Mathematical techniques of process-reconstruction to the underlying stochastic process, using coding and block entropies to characterize the measure and memory range, were applied. These modeling techniques allow a precise knowledge of the drive dynamics, fundamental topic in modern control approaches, without using difficult to obtain variables.