Abstract
Recently, radar high-resolution range profile (HRRP) recognition based on convolutional neural networks (CNNs) has received considerable attention due to its robustness to translation and amplitude changes. Most of the existing methods require that sufficient labeled data with complete aspect angles be used as training data, which is a difficult task in practice. In addition, HRRP signals have a high sensitivity to the aspect angle. Therefore, the representative and discriminative powers of the features extracted from typical CNN models are reduced due to incomplete aspect angles in the training data, which significantly limit the recognition performance. This paper first considers the problem of HRRP recognition with incomplete aspect angle training data and addresses the problem by a deep transfer learning framework. Specifically, the two proposed methods enhance the recognition performance by exploring the discriminative power and the intraclass consistency with auxiliary data, which have HRRP signals with complete aspect angles. This paper generates a simulated HRRP dataset from public data to validate the proposed work. The comparisons of the recognition results demonstrate that the proposed framework outperforms the latest CNN-based models.
Highlights
The high-resolution range profile (HRRP) is the vector sum of the radar echo projected on the ray, as shown in Figure 1, which is the sequence of the target scattering intensity distribution [1]
Target recognition based on HRRP signals plays an important role in radar automatic target recognition (RATR) systems for the following reasons: (1) the HRRP signal contains comprehensive information, including the target’s material, size, scattering information, and the electromagnetic characteristics of the target for radar-related applications; (2) compared to synthetic aperture radar (SAR) images, the HRRP is in vector form, which effectively reduces the storage capacity and computational cost; and (3) only the transmission radar wide-band signals are required to obtain the HRRP sequence [2], [3]
To improve the recognition performance of incomplete targets, this paper introduces inductive and transductive transfer learning
Summary
The high-resolution range profile (HRRP) is the vector sum of the radar echo projected on the ray, as shown in Figure 1, which is the sequence of the target scattering intensity distribution [1]. 2) Motivated by the application scenarios, the proposed work introduces an auxiliary dataset with complete aspect angle HRRP signals as the source domain. 3) The proposed work further considers the intraclass consistency in the auxiliary dataset by introducing a deep domain adaptation network to benefit the recognition task by reducing the sensitivity to the aspect angles of the HRRP signals. Unlike the definition in the previous stage, the HRRP signals from the captured and uncaptured aspect angles from complete targets are defined as the source and target domains, respectively. The intraclass or intratarget consistency is enforced by introducing a regularization term to the CNN network for feature extraction This term aims to minimize the distribution distances between the captured and uncaptured angle data features in the complete target to simultaneously emphasize the robustness and representative power of the extracted features.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
