ResNet Optimization Model for High-resolution Range Profile Target Recognition

This paper proposes a radar high-resolution range profile (HRRP) recognition algorithm based on ResNet, combined with the SE (Squeeze-and-Excitation) channel attention mechanism. In most current HRRP target recognition methods, feature extraction largely relies on manual extraction, followed by recognition through machine learning techniques. With the emergence of intelligent System-on-Chip (SoC) devices, which offer high-performance computing capabilities and low power consumption, it has become feasible to deploy deep learning algorithms on smaller radar platforms. By introducing the channel attention mechanism, the sensitivity and discriminative ability of the model to key features are enhanced, enabling ResNet to achieve high-precision recognition of different vehicle targets and corner reflector arrays. Validation using real-world data demonstrates that the proposed recognition algorithm yields superior results.