首页> 外文会议>Saint Petersburg International Conference on Integrated Navigation Systems; 20050523-25; Saint Petersburg(RU) >NEURAL NETWORK BASED TRAJECTORY METAMODEL FOR CONCEPTUAL DESIGN AND GUIDANCE OF LAUNCH VEHICLE
【24h】

NEURAL NETWORK BASED TRAJECTORY METAMODEL FOR CONCEPTUAL DESIGN AND GUIDANCE OF LAUNCH VEHICLE

机译:基于神经网络的弹道模型在概念设计和弹射制导中的应用

获取原文
获取原文并翻译 | 示例

摘要

This paper proposes a trajectory metamodel based on artificial neural network, firstly for integrated conceptual design and trajectory optimization and secondly for close loop guidance using onboard optimization of trajectory in near real time. Genetic Algorithm evaluates weight, propulsion, aerodynamics and flight dynamics to evolve to near optimal solution of minimum launch weight while meeting the constraints. Neural Network trajectory metamodel is used by Genetic Algorithm to rapidly find near global minimum. Sequential Quadratic Programming (SQP) starts from this initial guess and converges to local optimal solution. The most significant contribution of meta-modeling strategy is the drastic reduction in overall computation time, due to greatly reduced number of exact analysis required. In the guidance application, SQP uses onboard neural network trajectory metamodel to rapidly optimize the profile of angle of attack. Dispersion analysis of open loop and proposed closed loop guidance employing neural network are compared and shows the proposed scheme makes the system more robust against uncertainty in estimation of aerodynamic parameters.
机译:本文提出了一种基于人工神经网络的轨迹元模型,首先用于概念设计和轨迹优化的集成,其次用于基于机载轨迹的近实时优化的闭环引导。遗传算法评估重量,推进力,空气动力学和飞行动力学,以在满足约束条件的情况下发展为最小发射重量的最佳解决方案。遗传算法使用神经网络轨迹元模型快速找到接近全局最小值。顺序二次规划(SQP)从此初始猜测开始,并收敛到局部最优解。由于大大减少了所需的精确分析次数,因此元建模策略最重要的贡献就是大大减少了总体计算时间。在制导应用中,SQP使用机载神经网络轨迹元模型来快速优化攻角轮廓。比较了开环的色散分析和采用神经网络的闭环制导,并表明所提出的方案使系统对气动参数估计的不确定性更加鲁棒。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号