Functional link neural network-based intelligent sensors for harsh environments.

Abstract

As the use of sensors is wide spread, the need to develop intelligent sensors that can automatically carry out calibration, compensate for the nonlinearity and mitigate the undesirable influence of the environmental parameters, is obvious. Smart sensing is needed for accurate and reliable readout of the measurand, especially when the sensor is operating in harsh environments. Here, we propose a novel computationally-efficient functional link neural network (FLNN) that effectively linearizes the response characteristics, compensates for the nonidealities, and calibrates automatically. With an example of a capacitive pressure sensor and through extensive simulation studies, we have shown that the performance of the FLNN-based sensor model is similar to that of a multilayer perceptron (MLP)-based model although the former has much lower computational requirement. The FLNN model is capable of producing linearized readout of the applied pressure with a full-scale error of only ±1.0% over a wide operating range of −50 to 200 ˚C.