滑翔飞行器速度约束条件下制导律设计

Guidance Law Design for Glide Vehicles Under Velocity Constraint

针对滑翔飞行器速度约束条件下的制导问题,提出一种基于神经网络的制导方法。首先,建立滑翔飞行器弹道优化模型,给出基于高斯伪谱法的弹道优化方法和基于神经网络的制导律模型;其次,为进一步满足速度约束条件,提出了速度控制策略;最后,将高斯伪谱法获得的最优控制量离散后作为标签数据,利用反向传播(BP)神经网络进行训练,将训练好的网络模型嵌入弹道仿真环境中进行数值计算,并与定攻角剖面和高斯伪谱法离线优化的计算结果进行对比分析。仿真结果表明:提出的制导方法能够在满足速度约束条件下实现制导飞行,具备在线实施的能力和较好的工程适用性,对此类飞行器的制导律设计具有一定的参考意义。

A neural network-based guidance method is proposed for the guidance problem of glide vehicles under velocity constraints.First,a trajectory optimization model for glide vehicles is established,and the trajectory optimization method based on the Gauss pseudospectral method and the guidance law model based on the neural network are proposed.Afterwards,in order to further meet the velocity constraints,a strategy for velocity control is proposed.Finally,the optimal control quantity obtained by the Gauss pseudospectral method is discretized as the tag data,which are used to train the back propagation(BP)neural network.The trained network is then embedded into the simulation model for numerical calculation,and the obtained results are compared with those obtained by the fixed control profile and Gauss pseudospectral method.The simulation results show that the guidance method proposed in this paper can meet the velocity constraints.It has the ability of online implementation and good engineering applications,which have certain reference significance for the guidance law design for such glide vehicles.