摘要
防空导弹垂直发射转弯控制是一非线性最优控制问题。在优化设计垂直发射转弯规律时,粒子群算法存在易早熟而陷入停滞问题,而变尺度法对俯仰角规律初值敏感容易发散,且计算量大。利用粒子群算法无需手动设定初始值、速度快的特点,将其优化结果作为后者的初始值,然后利用变尺度法较强的局部搜索能力逼近最优解。采用这2种算法简单串联的方法优化得出的垂直发射转弯控制规律,经仿真验证可满足速度、攻角、过载等约束条件,俯仰角速度和攻角变化规律与工程经验一致。
Vertical launch control of air-defense missile was a non-linear optimization control problem.In the design of vertical launch turning control law,the PSO had the disadvantage of easily becoming premature and stagnation,while the BFGS was sensitive to the initial value of pitch angle law and was time consuming.To solve these problems the two methods were joined together: the PSO was used to get initial value for the BFGS as it was fast and didn't need to set initial value manually,then the BFGS's better local search ability was utilized to approach the optimization.The simulation results show that the optimized vertical launch control law satisfies the velocity,attack angle and overload constraints,and the pitch rate and attack angle curves are consistent with the engineering experience.
出处
《固体火箭技术》
EI
CAS
CSCD
北大核心
2012年第2期152-156,共5页
Journal of Solid Rocket Technology
关键词
垂直发射
弹道优化
粒子群优化算法
变尺度法
vertical launch
trajectory optimization
particle swarm optimization(PSO)
BFGS