针对机动目标的多约束末端三维制导控制研究

项目名称： 针对机动目标的多约束末端三维制导控制研究

项目编号： No.61304224

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 韩刚

作者单位： 中国科学院自动化研究所

项目金额： 26万元

中文摘要： 随着飞船交会对接、舰载无人机自主着落、防空反导等技术的发展，其共同核心难点是多约束复杂条件下目标高速随机机动时的制导控制问题。首先，针对目标机动可能造成探测信息不可靠等问题，基于估计制导一体化，提出了条件判别状态机动目标状态估计模型和运动建模方法；针对约束条件种类复杂、取值受环境变化影响的问题，提出了多约束条件集多源融合在线生成和离线修正方法，并研究一种落点、角度、速度和时间耦合约束下的自适应模糊神经网络滑模三维制导律算法。其次，为充分利用各种制导律之间的互补性，提出了一种基于目标运动模型、多约束条件集和多制导律模型库的多模融合制导方法。最后，针对传统仿真验证方法难以验证实际算法的复杂度与工程实现的可行度等问题，研究适用于三维制导控制实时并行计算的原型系统，软硬件协同实现算法的加速和性能优化。本项目对提高复杂条件下的制导精度和控制响应速度、提升智能自主制导控制水平具有重要的科学意义。

中文关键词： 机动目标状态估计；多约束三维制导律；滑模变结构控制；神经网络；分布式仿真计算

英文摘要： As the development of science and technology of the spacecraft rendezvous and docking, UAV autonomous landing on aircraft carrier, air defense and antimissile, and so on, Endgame guidance and control with Multiple Constraints against maneuvering targets is the common key difficult point. First，thinking about target moving maybe causing detecting data unreliable and in order to combine estimation and guidance, a maneuvering target state estimation and motion modeling method using observed data directly to estimate the target state based on conditional posterior probability is studied. As the types of constraints are complex and values effect by the application conditions, a method of multiple constrains set generated on-line through multi source information fusion and revised off-line is provided. Meanwhile, three-dimensional adaptive fuzzy neural network sliding mode guidance with multi-constraints of miss-distance, impact angle, velocity and time is developed. Second, a multi-model fusion guidance method based on target motion models, multi-constraints set and multiple guidance laws library is proposed in order to utility the complementary of various guidance law. Finally, traditionally it is difficult to simulate and validate the indeed complexity and practicability of the algorithms using simulation methods o

英文关键词： Maneuvering Target State Estimation；Guidance Law with Multiple Constraints；Sliding Mode Variable Structure Control；Neural Network Control；Distributed Simulation and Computing