| With the development and transformation of military science and technology,electronic warfare has become a major factor in the success or failure of modern warfare.As the precursor and cornerstone of electronic warfare,electronic reconnaissance is one of the important means of military intelligence reconnaissance,and its importance is self-evident.With the development of the current phased array,digital array and anti-jamming technology,the function of radar has changed from single to multi-function,and the modulation pattern has changed from single modulation to composite modulation,which undoubtedly increases the difficulty and complexity of radar working pattern recognition and intention reasoning and prediction Spend.Therefore,how to accurately and efficiently identify the working mode and behavior intention of radar has important theoretical significance and military values.This paper mainly takes the airborne active phased array radar as the research object,simulates the typical combat scene,respectively proposes a classification for the multiworking mode of the phased array radar and algorithms for recognition and intent inference predictions,which is based on the in-depth analysis of the different working modes of the phased array radar.The main focus of this paper is as follows:Firstly,starting from the system composition principle of the phased array radar,combined with the antenna scanning characteristics,wave position arrangement,resource management and scheduling of the phased array radar,its search,tracking and other working methods are analyzed,and the the typical combat scenarios of the phased array radar are simulated and modeled,and the characteristic parameters of the phased array radar in the time domain,frequency domain and air domain in different working modes are deeply analyzed,and the parameters that can characterize the characteristics of different working modes are extracted in the follow-up phased array radar work pattern recognition and intention reasoning prediction.Secondly,the identification algorithms of two phased array radar working modes are studied.First,by using the advantages of the K-nearest neighbor algorithm in the multi-classification problem and the 1-NN strategy in the blurred boundary elimination problem of the training data set,the multi-working mode of the phased array radar is classified and identified,and the signal-to-noise ratio is 10 d B,except that the recognition accuracy of TAS mode and SAM mode is low,the accuracy of other working modes can reach more than 90%;Second,the Trans H model in the knowledge graph is used to continuously vectorize the discrete feature information of the phased array radar under different working modes,and then the prior knowledge after the vector representation is used as the input of the MLP classification and recognition network in deep learning,so that the deep learning network can make full use of the prior knowledge in different working modes,and at the same time it can reduce the recognition model’s dependence on large-scale labeled data,and achieve a good recognition effect in the recognition of multiple working modes.Even when the SNR is 5d B,the recognition accuracy of various working modes can reach more than 95%.Finally,the prediction algorithm of phased array radar intent reasoning is studied.With the advantage of 1D-CNN network for feature extraction of time series,combined with Bi-GRU network with time memory function,an intent reasoning prediction algorithm in typical scenarios of phased array radar is proposed.After experimental comparison and analysis,the inference prediction performance based on the 1D-CNN-Bi-GRU network is better than the simple Bi-GRU network. |
