Radar Working Modes Recognition Based on Discrete Process Neural Network

Accurate identification of radar operation modes is the important premise of threat level assessment and interference decision. But parameters overlap of PDW between different radar working modes seriously affects the recognition accuracy. A discrete process neural network (DPNN) based on particle swarm optimization (PSO) training is proposed to realize radar working modes recognition. Firstly, radar syntactic modeling method is proposed to extract radar phrases as operation modes character description. Then, the appropriate DPNN structure is built to and trained via PSO. Finally, radar working mode recognition of unknown radar phrases is realized by the finished DPNN. Different from traditional machine learning method based on single sampling of radar signals, this method achieve recognition according to accumulation of radar pulse sequence, and make the best use of time series change law of radar signals. The simulation results show that compared with traditional machine learning method, such as LSSVM, BPNN, working modes recognition rate of the novel method increases significantly under the condition of serious parameter overlap.