基于改进动态时间规整的雷达状态识别方法

Airborne Radar State Recognition Based on PCA-DTW

针对机载雷达行为感知特征维度高,感知准确率低的问题,在典型雷达行为多级建模的基础上,将脉冲描述字(PDW)序列数据作为输入,提出一种基于语义模式距离的动态时间规整(DTW)雷达行为感知算法。该方法基于主成分分析(PCA)方法进行特征降维,并利用时间规整函数描述测试模板和参考模板的时间对应关系,求解两模板匹配时累计距离最小所对应的规整函数,据此进行机载雷达工作模式与状态的感知。仿真结果表明,该方法在较好的实时性前提下对机载雷达工作模式与状态的整体感知准确率达到90%以上,证明该方法有效可行,具有较好的工程应用前景。

Accurately perceiving the target s airborne radar behavior is the key to winning the air battle.Aiming at the high feature dimension and low accuracy of airborne radar behavior perception,this paper proposed a radar behavior perception algorithm based on dynamic time warping(DTW).On the basis of multi-level modeling of typical radar behavior,taking pulse description word(PDW)sequence data as input,a DTW sensing method based on semantic pattern distance was proposed.The feature dimension was reduced based on principal component analysis(PCA),and the time correlation between test template and reference template was described by time warping function to solve the regularization function corresponding to the minimum cumulative distance when the two templates were matched,and then perceived the working mode and state of airborne radar.Simulation results showed that the overall sensing accuracy of the method to the working mode and state of airborne radar under the premise of good real-time performance was more than 90%,which proved that the method was effective and feasible,and had a good engineering application prospect.