基于自适应模糊神经网络系统的射击毁伤评估模型研究

A Damage Assessment Model Based on Adaptive Neuro-fuzzy Inference System

针对炮兵对集群有生力量目标射击的毁伤效果评估问题,提出一种基于自适应模糊神经网络系统(ANFIS)的评估新模型。应用坐标毁伤律仿真计算不同炸点分布时的毁伤效果,形成样本数据,采用ANFIS学习出毁伤因素与毁伤效果之间的映射关系,进而对毁伤效果进行评估。运用该模型对炮兵连射击毁伤效果进行评估的实验结果表明,评估值和实际值之差约在3%左右,充分说明该模型精度高,泛化能力好,具有很强实用性,能有效根据弹群信息进行毁伤评估。

A novel model for predicting artillery firing damage effect to group target based on adaptive neuro-fuzzy inference system （ANFIS） was presented. After damage probabilities under different firing conditions were calculated based on law of burst coordinate, an ANFIS was trained to learn the mapping model between damage factors and probability. It was then used to assess the firing damage. Experiment results show that the differences between the outputs of ANFIS and simulation results are about 3% or so, and it indicates that the model has satisfying accuracy and extensive ability.