具不确定观测和相关噪声的最优递推滤波

应用正交投影定理,对于噪声相关系统,在观测不确定的情况下,求出了状态向量的卡尔曼滤波公式。在线性最小方差准则下,解决了状态噪声一步相关,相同时刻的测量噪声和状态噪声相关,以及测量噪声和状态噪声在相邻时刻均相关的情况下的滤波问题,并给出了递推滤波公式。仿真算例说明了它们的有效性。     

A novel control method for turboshaft engine with variable rotor speed based on the Ngdot estimator through LQG/LTR and rotor predicted torque feedforward

In order to compensate for the disturbance of wide variation in rotor demanded torque on power turbine speed and realize the fast response control of turboshaft engine during variable rotor speed, a cascade PID control method based on the acceleration estimator of gas turbine speed (Ngdot) and rotor predicted torque feedforward is proposed. Firstly, a two-speed Dual Clutch Transmission (DCT) model is applied in the integrated rotor/turboshaft engine system to achieve variable rotor speed. Then, an online estimation method of Ngdot based on the Linear Quadratic Gaussian with Loop Transfer Recovery (LQG/LTR) is proposed for power turbine speed cascade control. Finally, according to the cascade PID controller based on Ngdot estimator, a rotor demanded torque predicted method based on the Min-batch Gradient Descent-Neural Network (MGD-NN) is put forward to compromise the influence of rotor torque interference. The simulation results show that compared with cascade PID controller based on Ngdot estimator and the one combined with collective pitch feedforward control, the novel control method proposed can reduce the overshoot of power turbine speed by more than 20%, which possesses faster response, superior dynamic effect and satisfactory robustness performance. The control method proposed can realize the fast response control of turboshaft engine with variable rotor speed better.     

基于决策树的涡轴发动机气路健康评估方法

为解决涡轴发动机退化模式多、故障样本大、分类困难的问题,将梯度提升决策树(GB?DT)算法应用于涡轴发动机气路健康评估.首先对涡轴发动机故障标签进行了拆解,针对每一类标签设计了一个分类器.通过特征重要度排序筛选了6个分类标签的输入特征,有效降低模型复杂度.通过遍历确定树的最优数量和深度,确保了分类的准确性.仿真结果表明...     

Relevant experience learning: A deep reinforcement learning method for UAV autonomous motion planning in complex unknown environments

Unmanned Aerial Vehicles (UAVs) play a vital role in military warfare. In a variety of battlefield mission scenarios, UAVs are required to safely fly to designated locations without human intervention. Therefore, finding a suitable method to solve the UAV Autonomous Motion Planning (AMP) problem can improve the success rate of UAV missions to a certain extent. In recent years, many studies have used Deep Reinforcement Learning (DRL) methods to address the AMP problem and have achieved good results. From the perspective of sampling, this paper designs a sampling method with double-screening, combines it with the Deep Deterministic Policy Gradient (DDPG) algorithm, and proposes the Relevant Experience Learning-DDPG (REL-DDPG) algorithm. The REL-DDPG algorithm uses a Prioritized Experience Replay (PER) mechanism to break the correlation of continuous experiences in the experience pool, finds the experiences most similar to the current state to learn according to the theory in human education, and expands the influence of the learning process on action selection at the current state. All experiments are applied in a complex unknown simulation environment constructed based on the parameters of a real UAV. The training experiments show that REL-DDPG improves the convergence speed and the convergence result compared to the state-of-the-art DDPG algorithm, while the testing experiments show the applicability of the algorithm and investigate the performance under different parameter conditions.     

基于连续蚁群算法的Bayesian方位估计快速方法

针对蚁群算法(ACO)在解决高维非线性搜索问题方面的有效性,提出了基于蚁群优化算法的Bayesian最大后验概率方位估计(ACO-Bayesian)快速方法.该方法将Bayesian最大后验概率函数作为蚁群算法的目标函数,选取若干一维高斯函数的加权和作为连续蚁群算法中信息量概率分布函数,经过有限次迭代得到Bayesian方法的非线性全局最优解.仿真结果表明,ACO-Bayesian方法在保持Bayesian方法优良性能的同时,将Bayesian方法的计算量减少到原来的1/14.水池实验结果验证了ACO-Bayesian方法的正确性和有效性,为其工程应用奠定了基础.      

基于多源卫星融合资料的中国近海海洋锋检测研究

海洋锋是海洋动力学中重要的研究对象.目前主要利用卫星遥感海面温度数据分析海洋锋.由于夏季中国近海海域海面温度趋同特性,无法利用海面温度数据分析海洋锋,而卫星遥感海色数据可以有效弥补海面温度检测海洋锋的不足,提高海洋锋检测的准确性.基于多源卫星融合数据,利用梯度检测算法对中国近海海域进行海洋锋研究,结合海面温度和蓝绿波段...     

Improvement of Jahangir＇ s multiple moments estimator

在M.Jahangir以常数为权的组合式矩估计器的基础上,给出一种以函数为权的组合式矩估计器,称为L-J估计器.其中,最优加权函数是根据U估计器与形状参数的单调关系,通过数论网格最优化算法搜索解出.大量仿真实验证实,在对K分布形状参数v大范围的参数估计中,L-J估计器在估计精度上,不但较Jahangir等提出的常数加权组合矩估计器的精度有显著提高,而且可与MLE（Maximum Likelihood Estimator）相当.特别是由于MLE作为渐进无偏估计量,需要充分大的样本长度才能达到最优,这就使得L-J估计器的估计精度可在样本长度较小时优于MLE.此外,L-J估计器无需迭代运算,因而在计算效率上,显著优于现有的ML估计器.     

改进的Sage—Husa滤波及在卫星姿态估计中的应用

基于极大后验估计原理,提出了一种改进的噪声估计器,以实现对噪声均值和方差的在线估计,抑制滤波器发散。对自适应扩展卡尔曼滤波算法在卫星姿态确定系统中的应用进行了仿真。结果表明新算法滤波精度优于扩展卡尔曼滤波（EKF）,与Sage—Husa自适应滤波算法相比,可阻止滤波器发散,提高系统滤波精度。     

一种地球紫外中心快速提取方法

针对地球紫外中心指向高精度提取问题,提出一种基于梯度统计的快速、低存储需求的紫外地平圆盘中心提取算法。首先,考虑到星载计算机与高内存消耗的滤波算法、实时导航需求的冲突,采用结合Sobel边缘算子与局部二值模式（LBP）算子的改进边缘快速提取方法,有效剔除背景噪声并准确提取地球紫外临边特征。然后对非完整临边边缘采用最小二乘拟合得到中心的精确位置。实验结果表明,该方法抗噪声,可实现亚像素级地球中心提取,并显著降低存储需求和计算时间开销。对于1046×746的图像,该算法需要的存储空间仅为1100kByte,运算时间在20ms内,满足自主导航的需求。     

挠性充液航天器超近程逼近段动力学建模与控制

超近程逼近过程中,服务航天器为刚柔液耦合的复杂系统,单摆等液体晃动模型不再适用,对这种工况下的航天器进行了动力学建模与控制的研究。采用虚功率原理推导了一种新的适用于三轴推力作用下的液体晃动等效模型,通过引入相对位置导引矢量与相对位置误差矢量,建立了相对轨道误差动力学模型,结合相对姿态动力学模型,得到超近程逼近段的刚柔液耦合的相对轨道姿态动力学模型。针对模型存在的不确定性和未知干扰,设计了基于滑模估计器的相对姿轨耦合控制律。通过数学仿真验证了控制律的有效性,仿真中的晃动结果与Flow-3D结果能够吻合,验证了晃动模型的合理性。