查看文章

This paper, after a brief theoretical background to Artificial Neural Networks (ANNs) reviews their utilization in the field of measurements. The criteria used for choosing between the possible ANN software and hardware implementations are then discussed. The advantages and disadvantages of the possible analogue or digital hardware implementation are also detailed. Some of the applications of ANNs in the measurements field, variously utilizing software, analogue or digital hardware implementations are shown. In particular, the strategies for building suitable ANN-based software models of mixed analogue/digital measurement devices are presented. Successively, a method for designing a neural network able to pre-process signals in an analogue mode is described. Finally, the architecture of a DSP-based neural apparatus for the real-time measurement of the components of a dual-tone multi-frequency signal …