Filtered forward-backward time-stepping method without information on incident field

A new forward-backward time-stepping (FBTS) method using gradient-based optimization method has been proposed for inverse scattering problems. This method is based on the field equivalence principle and does not require the explicit knowledge of the incident field. Due to the nonlinearity of the inverse scattering problem, the solution based on gradient methods often gets trapped into a local minimum. In order to avoid this problem, the measured total field data is processed through a low-pass filter with several cut-off frequency sequences in the FBTS algorithm. Numerical simulations show the effectiveness of the proposed method.