(A) study on the improvement of coil alignment in wireless power transmission system using sensor coils and recurrent neural network position estimation techniques

Abstract

Using a recurrent neural network (RNN), the past position of a moving vehicle can be considered for position localization to achieve higher accuracy compared to single-instance methods. We proposed the use of simple and low-cost sensor coils mounted on the receiver for alignment of wireless power transfer coils in parking vehicle application with RNN as the localization algorithm. A basic LSTM network was built and tuned to estimate the receiving coil’s position with the induced voltages on sensing coils as inputs. Our proposed system outperforms previously proposed method and may be quickly adapted and implemented to actual production system due to its simplicity.