Evaluation Algorithms for Linear Position Sensors Assisted by Artificial Neural Network

Abstract

An artificial neural network assisted compact linear inductive position sensor is presented. The sensor operates with a long primary printed circuit coil excited with periodically alternating current, and with multiple secondary printed circuit coils possessing different induced voltages as a position of a metallic or resonant target moving along the coil system. The secondary coil signals are phase sensitively demodulated, digitized and passed to the input layer of an artificial neural network trained to estimate the position of the target, its distance from the sensor and the signal quality.