An Intelligent Discrimination Method for Track Deception Jamming

For radar active deceptive jamming, a deep learning method for radar signal track discrimination is proposed based on end-to-end data processing mode, which can fuse target track information detected by multiple radar architecture. Meanwhile, the method combines convolutional neural network, bidirectional long and short-term memory network with attention mechanism, which can effectively extract multidimensional features of track while adaptively adjusting the weights of attention regions. Compared with the traditional point information processing method, using multidimensional track features can analyze the motion pattern and spatial characteristics for the target more comprehensive, leading to a better identification effect. Simulation result shows that the proposed method can be effectively applied to both active and passive radar architecture, and the discrimination effect of jamming based on track can reach over 99%. The discrimination ability and generalization performance of radar system are effectively improved. It will have high application value in information fusion center of multi-architecture radar on complex battlefields, which provides a new solution for practical implementation in the field of deception jamming countermeasures.