Modeling of an intelligent pressure sensor using functional link artificial neural networks.

Area of Science:

• * Electrical Engineering
• * Artificial Intelligence
• * Sensor Technology

Background:

• * Capacitor pressure sensors (CPS) require accurate modeling for reliable pressure readout.
• * Traditional modeling approaches may lack computational efficiency or robustness across varying conditions.
• * Artificial neural networks (ANNs) offer potential for complex nonlinear system modeling.

Purpose of the Study:

• * To develop a computationally efficient and robust intelligent model for capacitor pressure sensors (CPS).
• * To investigate the performance of a functional link artificial neural network (FLANN) for CPS modeling.
• * To compare the efficacy of different polynomial expansions within the FLANN architecture.

Main Methods:

• * Development of a functional link artificial neural network (FLANN) incorporating functional expansion units (Chebyshev, Legendre, power series).
• * Modeling of capacitor pressure sensor (CPS) behavior using the proposed FLANN.
• * Comparative analysis of FLANN models against a multilayer perceptron (MLP) model via computer simulation.

Main Results:

• * The FLANN demonstrated capability for complex nonlinear mapping between sensor input and output.
• * FLANN models achieved a maximum pressure estimation error within +/- 3% across a temperature range of -50 to 150 degrees C.
• * FLANN exhibited computational advantages over MLP for comparable modeling performance.

Conclusions:

• * The FLANN provides an accurate, computationally efficient, and robust method for modeling capacitor pressure sensors.
• * FLANN-based CPS models are suitable for economical and reliable implementation.
• * The study validates the effectiveness of FLANNs, particularly with polynomial functional expansions, for sensor modeling applications.