An Intelligent Target Feature Recognition Method Based on Multi-mode OAM Beams

Abstract

Target detection based on space modulation requires a large number of test modes with space-time independence. The Orbital Angular Momentum (OAM) beams are orthogonal to each other and have infinite modes. Due to the strong dispersive materials, multi-mode OAM beams with the same scattering angle can be generated in the frequency domain. In this manuscript, the propagation characteristics of multi-mode OAM beams are analyzed, which can be utilized to improve detection efficiency. The echoes from the target illuminated by the multi-mode OAM beams are then investigated in three different application scenarios. A convolution neural network is employed to extract the relationship between the echo data and the target image based on prior knowledge. The target and the imaging scenarios can be distinguished with a high probability. Finally, the proposed method’s anti-noise performance is analyzed. The experimental results show that in the ideal state, the accuracy of target scene judgment can reach 97.5%. The accuracy of the target location recognition is higher than 80% when the interval between two adjacent targets in a scene is larger than a threshold. The accuracy of the target location recognition in three scenes is greatly reduced when SNR is less than 20 dB, depending on the scene.

Keywords

• spatial modulation
• orbital angular momentum (oam)
• multi-mode vortex electromagnetic wave
• convolution neural network (cnn)
• feature extraction
• target recognition