二级入轨空天飞机上升轨迹优化

以最小燃料消耗量为性能指标,考虑了多种约束条件(过载、动压、加热率等 ),采用了组合优化方法对空天飞机上升轨迹进行优化。该组合优化算法包括静态参数与动 态参数优化方法。其中,静态参数优化算法采用了全范围搜索方法,动态参数优化算法包 括动态规划法与共轭梯度法。通过数值仿真计算,得到飞行分离状态数据及相应的最 优 上升轨迹。数值仿真表明,该组合算法对航天器轨迹优化具有良好的适用性及实际工程价 值。
     

基于Gauss伪谱法的高超声速飞行器多约束三维再入轨迹优化

高超声速飞行器再入飞行过程中,需要满足多种过程约束和终端状态约束,同时再入初始状态根据飞行任务不同会有较大变化,针对其特点的快速轨迹优化问题已成为当今热点。本文研究了一种基于"初值轨迹生成+Gauss伪谱法+SQP求解NLP"的方法,既利用了Gauss伪谱法收敛快、精度高的特点,又结合初值轨迹生成算法,弥补了Gauss伪谱法对初值敏感的不足。本文在仿真过程中选取再入总吸热量最小为性能指标,求解了满足多种约束的再入轨迹,并将优化的结果与数值积分的结果进行比较,验证了此算法有效性和可行性。     

基于轨迹线性化方法的导弹动态逆控制系统设计

基于动态逆控制及轨迹线性化方法,研究了一种新的导弹鲁棒自适应控制系统设计方法.首先,基于导弹的非线性模型设计快、较慢回路的动态逆控制器,以实现其非线性解耦;然后,在较慢回路中设计鲁棒自适应轨迹线性化控制器,以补偿整个控制系统的不确定性及干扰,即在轨迹线性化控制的基础上,通过设计RBF神经网络在线估计系统中存在的不确定性...     

Optimal solar sail trajectory approximation with finite Fourier series

The heliocentric transfer of a solar sail-based spacecraft is usually studied from an optimal perspective, by looking for the control law that minimizes the total flight time. The optimal control problem can be solved either with an indirect approach, whose solution is difficult to obtain due to its sensitivity to an initial guess of the costates, or with a direct method, which requires a good estimate of a feasible (guess) trajectory. This work presents a procedure to generate an approximate optimal trajectory through a finite Fourier series. The minimum time problem is solved using a nonlinear programming solver, in which the optimization parameters are the coefficients of the Fourier series and the positions of the spacecraft along the initial and target orbits. Suitable constraints are enforced on the direction and magnitude of the sail propulsive acceleration vector in order to obtain feasible solutions. A comparison with the numerical results from an indirect approach shows that the proposed method provides a good approximation of the optimal trajectory with a small computational effort.     

The Effects of Transverse Magnetic Field and Density Variations on the Particle Energy Spectra in a Reconnecting 3D Current Sheet

Electron and proton acceleration by a super-Dreicer electric field is further investigated in a non-neutral reconnecting current sheet (RCS) with a variable plasma density. The tangential B _z and transverse magnetic field components B _x are assumed to vary with the distances x and z from the X nullpoint linearly and exponentially, respectively; the longitudinal component (a ‘guiding field’) is accepted constant. Particles are found to gain a bulk of their energy in a thin region close to the X nullpoint where the RCS density increases with z exponentially with the index λ and the tangential magnetic field B _x also increases with z exponentially with the index α. For the RCS with a constant density (λ = 0), the variations of the tangential magnetic field lead to particle power-law energy spectra with the spectral indices γ₁ being dependent on the exponent α as: for protons and for electrons in a strong guiding field (β > 10⁻²) and for electrons in a moderate or weak guiding field (β > 10⁻⁴). For the RCS with an exponential density increase in the vicinity of the X nullpoint (λ≥ 0) there is a further increase of the resulting spectral indices γ that depends on the density exponent index λ as for protons and for electrons in weaker guiding fields and as for electrons in stronger guiding fields. These dependencies can explain a wide variety (1.5–10) of particle spectral indices observed in solar flares by the variations of a magnetic field topology and physical conditions in a reconnecting region. This can be used as a diagnostic tool for the investigation of the RCS dynamics from the accelerated particle spectra found from hard X-ray and microwave emission.     

蒙皮拉形加载轨迹优化

对双曲度飞机蒙皮拉伸成形问题,探讨了加载轨迹对成形质量的影响,提出了采 用最优化理论与算法和有限元数值模拟相结合,确定最优拉形加载轨迹的原理与方法.分析 了单凸双曲度蒙皮拉伸成形常用加载轨迹的成形特点和不足,采用有限元数值模拟和序列二 次规划相结合的方法,通过正交试验设计建立近似响应模型和设计变量的取值范围,以减小 卸载回弹提高贴模度为研究目标、最大主应变为约束条件建立最优化数学模型,确定最合理 的拉形加载轨迹.优化结果表明成形零件贴模度和应变分布与厚度的均匀程度均有明显提高 .      

基于改进粒子群算法辨识Volterra级数的目标机动轨迹预测

目标机动轨迹预测是空战态势感知和目标威胁评估的重要前提。针对传统目标机动轨迹预测模型复杂度大、预测精度低等问题，结合目标机动轨迹时间序列的混沌特性，提出一种基于相空间重构理论和Volterra泛函级数的目标机动轨迹预测模型。该模型首先采用0-1检测法验证了目标机动轨迹时间序列具有混沌特性；其次，利用C-C法确定嵌入维数和时间延迟，对目标机动轨迹时间序列进行了相空间重构；然后，引入Volterra泛函级数预测模型，为了克服高阶Volterra核函数求解复杂的难题，提出一种混沌变异自适应粒子群算法，构建一种基于改进粒子群算法辨识的Volterra级数预测模型，并将其应用于目标机动轨迹预测；最后，将所提算法与卡尔曼滤波算法以及机器学习算法进行单步和多步预测对比，同时将改进粒子群算法与其他智能算法进行性能比较。仿真结果表明：所提预测模型具有良好的单步和多步预测性能，改进的粒子群算法具有参数辨识精度高、收敛速度快的优点。     

集群无人机队形重构及虚拟仿真验证

队形重构是集群无人机（UAV）控制的重要问题，指无人机按照要求安全、无碰撞地从一个队形变换到另一个队形，其难点在于快速规划最优安全轨迹并控制无人机进行轨迹姿态的高精度跟踪。针对集群无人机队形重构的上述问题，首先，基于CAPT（Concurrent Assignment and Planning of Trajectories）算法，解决了多无人机的目标分配和轨迹生成的实时性问题，实现了集群无人机的最优安全路径规划；其次，提出一种有限时间多变量积分滑模连续控制算法，解决了无人机轨迹姿态的高精度跟踪问题，并通过MATLAB仿真验证了该控制算法的有效性；最后，为了更加真实直观地演示无人机三维仿真效果，建立了基于Gazebo-ROS的无人机仿真平台，实现了12架四旋翼无人机队形重构"建模-仿真-可视化"的一体化仿真演示，验证了上述路径规划算法和轨迹姿态控制算法的有效性。     

具有时间约束的无人机遗传算法航迹规划

为无人机进行航迹规划不仅要满足路径的有效性,还需要考虑无人机的到达时间约束.将遗传算法应用于动态环境中无人机的航迹规划,通过编码生成初始航迹,确定合理的适应度函数和进化操作,同时根据要求到达时间约束为每个航迹加入过点速度和到达时间信息,最终搜索规划出可飞且满足时间约束的航迹.在仿真实验中实现了针对移动目标的规划和突然出现威胁的局部重规划.