R. Belu, A. Paun, A. R. Belu
Neural Networks in Measurement, Instrumentation and Control

Abstract.
Artificial intelligence techniques are now attractive tools used to enhance and improve the efficiency, the capability, and the features of the instrumentation in application areas related to measurements, system identification, and control. These techniques exploit the computational capabilities of modern computing systems to manipulate the sampled input signals and extract desired measurements. The aim of this paper is to introduce the fundamental concepts of artificial neural networks and to present a review of those applications related to measurement, instrumentation and control.
Artificial neural networks have several advantages in practice, including learning and adapting ability, parallel distributed computation, robustness, etc. The use of artificial neural networks in instrumentation, measurement, and control applications has been increasing dramatically in the last three decades. This paper describes some of the neural network techniques providing the reader with in-depth theoretical elements and references to allow them to explore the applicability of artificial neural network technologies for their specific applications. One of the main topics of this paper will refer to surveying the neural signal understanding and classification methods, and the selected implementation and architecture will be described with its performance in terms of correct classification rates and robustness to noise. The chosen forms of neural networks were tested using different types of input information. This paper is the first in a series of three papers dealing with problems related with applications of neural artificial networks in measurement, instrumentation and control.