针对机动目标的三维实时滚动优化制导策略

为解决目标机动策略未知条件下的飞行器拦截问题, 提出一种基于神经网络的三维滚动优化制导策略。首先, 针对全局最优导引律终端时刻难以确定的问题, 在滚动时域优化框架下, 引入零效脱靶量设计局部最优导引律, 并使用粒子群优化算法进行求解。其次, 为了提高制导律在线求解效率, 构建神经网络, 对优化算法滚动求解得到的若干组制导训练数据进行离线学习, 并将经过训练的网络用于制导指令在线滚动优化。仿真结果表明, 神经网络-滚动优化制导策略对采取各类机动方式的目标均具有较好的制导性能, 有效提高了制导指令在线优化效率, 可以为飞行器制导律实时滚动求解提供参考。

Three dimensional real-time rolling optimal guidance strategy for maneuvering targets

In order to solve the interception problem of aircraft under the condition of unknown target maneuver strategy, a three-dimensional rolling optimal guidance strategy based on neural network is proposed. Firstly, aiming at the problem that it is difficult to determine the terminal time of the global optimal guidance law, the zero effort miss is introduced to design the local optimal guidance law under the framework of rolling horizon optimization, and the particle swarm optimization algorithm is used to solve it. Secondly, in order to improve the efficiency of on-line solution of guidance law, a neural network is constructed to learn several groups of guidance training data obtained by rolling solution of particle swarm optimization algorithm off-line, and the trained network is used for on-line rolling optimization of guidance instructions. The simulation results show that the neural network rolling optimization guidance strategy has good guidance performance for targets with various maneuver modes, effectively improves the on-line optimization efficiency of guidance instructions, and can provide a reference for the real-time rolling solution of aircraft guidance law.