A Novel Method for Eddy Current based Velocity Estimation by Magnetostrictive Position Sensors

Magnetostrictive position sensors (MPS) are used for absolute long-range as well as for high-precision linear position measurement. The working principle of these sensors is based on a time-of-flight (TOF) measurement of a structure-borne sound wave within a magnetostrictive waveguide. State-of-the-art techniques only determine the position based on the measured TOF. Physically related parameters like velocity or acceleration are calculated by derivation of the determined position. In this paper, a novel method for real-time simultaneous estimation of position and velocity of a position marker is presented. Related to the position markers velocity eddy currents are induced within the MPS housing. The influence of the magnetic fields generated by this eddy currents leads to a velocity dependent impact on the structure-borne sound wave that is detected and evaluated by means of an artificial neural network. The training dataset for the velocity estimation is created from recorded datasets itself, without an external velocity reference system (self-training). Therefore, the electrical representation of the detected structure-borne sound wave is digitalized at high speed. Using the presented method, the velocity data could be used either as an additional independent parameter for consistency checking (self-diagnostics), or for the improvement of the position measurement by signal fusion (self-optimization).