Artificial Neural Network Assisted Compact Inductive Distance Sensor

A material independent compact industrial linear inductive distance sensor is presented. The sensor operates with a gradiometric coil arrangement and pulsed excitation. Depending on the distance and the material quality of the metallic object to be sensed, the transient pre-amplified induced voltage signal may possess various magnitude and time course, which is further complicated by the influence of the constructional parts of the sensor (e.g. shielding). The time dependent transient signal waveforms are evaluated by artificial neural networks in terms of target position, and a compact sensor with sufficiently high update rate has been designed with traditional industrial analog outputs and IO-Link interface.