Shadow Region Imaging Algorithm with Aperture Synthesis of Multiple Scattered Waves for UWB Radars Shouhei KIDERA ∗ , Takuya SAKAMOTO and Toru SATO Graduate School of Informatics, Kyoto University, Japan Email:kidera@aso.cce.i.kyoto-u.ac.jp 1 Introduction Ultra-wide band (UWB) pulse radar is promising as a near field sensing technique with high range resolution. As such, it is applicable to non-contact measurement of precision devices with specular surfaces, or to security systems that can identify a human body in invisible situations. For these applications, the SAR (Synthetic Aperture Radar) algorithm is still promising, as it creates a stable and accurate target image even in the near field [1]. However, in the case of complex or multiple targets, this algorithm suffers from increased shadow regions or false images caused by multiple scattered waves. In most case, a multiple scattered wave propagates a different path from that of a single scattered wave. This means that the multiple scattered echo has independent information on target surfaces, and thus has the potential to improve the image quality of the conventional methods, which use only single scattered waves. Although the time reversal algorithms with multiple scat- tered waves have been proposed, when focusing on target detection or positioning in cluttered situations [2-4], these require a target modeling or a priori information of the surrounding environment such as the walls. To relax these conditions, this paper proposes a direct imaging algorithm based on aperture synthesis of multiple scattered echoes. As a novelty of this paper, the proposed method is applicable to arbitrary target shapes, and directly enlarges the visible range on the target surface. Results obtained from numerical simulation verify the effectiveness of the proposed method. 2 System Model Fig. 1 illustrates the system model, which assumes that the target has high con- ductivity such as a metal, and an arbitrary shape with a clear boundary. The propagation speed of the radio wave c is assumed to be known constant. An omni- directional antenna is scanned along the x-axis. We use a mono-cycle pulse as the transmitting current. The real space in which the target and antenna are located is expressed by the parameters r =(x, z), which are normalized by λ, the central wavelength of the pulse. s(X, Z ) is defined as the output of the Wiener filter at the antenna location (x, z )=(X, 0), where Z =ct/(2λ) is expressed by the time t. 3 Conventional Algorithm The SAR algorithm can create a stable and accurate target image even in the near field. The distribution image I 1 (r) obtained with this algorithm is calculated as I 1 (r)=  X∈Γ s  X,  (x - X ) 2 + z 2  dX, (1)