Airborne Radar Operating State Recognition Based on Multi-Parameter Fusion RCS-Conv Method

Airborne radar operating state directly reflects the combat intention of the target platform, playing a crucial role in electronic warfare. However, the challenge of parameter overlap among different operating states poses a new obstacle to radar state recognition. To enhance the capability of recognizing radar operating states with overlapping parameters, a multi-parameter fusion residual coordinate shrinkage ConvNet (RCS-Conv) is proposed in this paper. First, the temporal regularities of electromagnetic data and the correlation features among parameters are analyzed. Subsequently, discriminative features conducive to radar state recognition are identified. Moreover, a stream-level deep recognition framework incorporating a coordinate attention (CA) mechanism is meticulously designed. This recognition method not only captures inter-channel dependencies and positional information but also extracts discriminative timing and parameter correlation features. Simulation results demonstrate that the proposed method achieves accurate recognition of airborne radar states in complex electromagnetic environments.