Foreign Object Detection for Wireless Power Transfer Systems Using MLMVN

In this paper, a technique for detecting foreign metal objects in a wireless power transfer system is presented. The proposed approach exploits a multilayer neural network with multivalued neurons. The main advantage of the proposed approach compared to the literature is the reduced number of current and voltage measurements. The procedure proposed in this work is completely based on real measurements, which are interpolated to obtain a suitable training dataset. The detection results are comparable to those of other techniques available in the literature, which do not require the introduction of additional coils.