Radar Antenna Scanning Pattern Recognition Method Based on LSTM Network

Radar antenna scanning patterns include mechanical scanning, one-dimensional phased array scanning, and two-dimensional phased array scanning. When facing radar emitter whose radar type is unknown, the antenna scanning pattern is a critical factor in assessing the threat level. Traditional identification methods are based on expert features and often fail to adequately consider factors such as amplitude-frequency response and measurement errors. This paper employs short-term interpolation for noise reduction and long-term LSTM for classification and recognition. Experiments show that this method has good accuracy in identifying pulse sequences under conditions of frequency variation and measurement errors, making it valuable for practical applications.