基于BP神经网络的TBM拦截效果评估

为了综合考虑TBM拦截作战中各种因素对拦截效果评估的影响,基于红外成像、ISAR成像和机动目标跟踪3种方法分析了TBM拦截效果评估体系的构成,利用神经网络在处理非线性复杂问题上的优势,提出了基于BP神经网络的TBM拦截效果评估模型,并详细阐述了模型的构建过程。由于标准BP算法存在易形成局部极小点和收敛速度慢等问题,采用加入动量项和实时调整系数法对算法进行了改进。仿真分析证明：改进的BP算法具有较快的收敛速度和较高的收敛精度,同时验证了该神经网络模型在TBM拦截效果评估中的有效性。

TBM Intercepting Effect Assessment Based on BP Neural Network

TBM intercepting effect assessment is an important and complex job in the TBMD. In order to comprehensively consider the influence of various factors on the intercepting effect assessment in TBM intercepting combat, the assessment system is analyzed based on infrared imaging, ISAR imaging and maneuvering target tracking methods. In view of Neural Network'' advantage in dealing with the complex problems, an assessment model with BP Neural Network is built, and the building process is discussed in detail. The standard BP algorithm with the problems that the convergence speed is slow and local minimum points are easily formed, is improved through adding momentum top and adjusting factors timely. Finally, the improved BP algorithm is simulated through living example and analyzed, the result shows that the improved algorithm is better in convergence speed and accuracy, and simultaneously verifies the validity and reliability of the model in TBM intercepting effect assessment.