High-order maneuvering target detection based on deep neural network
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
High-order maneuvering targets are often accompanied by range migration and Doppler-frequency migration, and exhaustive searches over high-dimensional parameter spaces impose a heavy computational burden. To address these issues, a pyramid residual network (PRNet) with large-kernel convolutions is proposed. Firstly, a nonlinear mapping between radar echoes and higher-order motion trajectories is established through PRNet, where large-kernel convolutions enable multiscale feature extraction, allowing the network to infer target trajectories from noisy backgrounds. Secondly, instead of performing exhaustive searches over the full parameter space, the proposed method searches only within a candidate subspace, thereby reducing computational cost. Finally, a visualization technique is incorporated to enhance the interpretability of the proposed network and provide more intuitive insights. Experimental results demonstrate that the proposed method outperforms conventional algorithms in both detection accuracy and computational efficiency.
Fields of Study & Subjects
Radar Systems and Signal Processing
Advanced SAR Imaging Techniques
Advanced Neural Network Applications
Publication details
- Journal / source
- Journal of Systems Engineering and Electronics
- Pages
- 1–12
- Publication date
- 1 January 2026
- Language
- EN