机动发射弹道导弹集群诸元快速规划

针对机动发射条件下弹道导弹集群的飞行诸元快速规划问题，将神经网络预测与最小二乘优化相结合，提出了一种弹道导弹发射诸元快速规划方法。首先分析了弹道导弹助推段飞行策略并选取适当的发射诸元，以发落点信息为输入，设计双隐藏层诸元预测网络，通过弹道仿真获取弹道数据建立数据集完成网络训练，利用该网络可以得到发射诸元迭代初值。在此基础上，为了消除数据集中样本数据不平衡对发射诸元规划精度的影响，以落点射程、横程、高程偏差最小为指标函数，结合最小二乘优化方法进行迭代获得发射诸元精确解。最后在典型发射场景下，进行了弹道导弹集群机动快速发射仿真验证。结果表明，该方法相较于传统方法可显著提高计算速度与精度，且在给定的大范围机动条件下，能够满足弹道导弹集群对远距离、多目标的快速精确打击。

Rapid Data Planning of Mobile Launched Ballistic Missile Cluster

Aiming at the rapid planning of flight data of ballistic missile cluster under the condition of mobile launch, a rapid planning method of ballistic missile launch data is proposed by combining neural network prediction with least square optimization. Firstly, the flight strategy of the ballistic missile in the boost phase is analyzed and the appropriate launch data is selected. With the launch and landing point information as the inputs, the dual hidden layer data prediction network is designed. The trajectory data is obtained through the trajectory simulation, and the dataset is established to complete the network training. The iterative initial value of the launch data can be obtained by using the network. On this basis, in order to eliminate the influence of imbalance of sample data in the data set on the planning accuracy of launch data, the minimum deviation of range, cross range and elevation of the landing point is taken as the index function, and the accurate solution of launch data is obtained by iteration with the least square optimization method. Finally, the rapid mobile launch simulation of ballistic missile cluster is carried out in a typical launch scenario. The results show that this method can significantly improve the computing speed and accuracy compared with the traditional method, and can meet the rapid and accurate attack of ballistic missile cluster against the long distance and multiple targets under the given large range of maneuvering conditions.