重访机制驱动的多无人机协同动目标搜索方法

为实现多无人机高效捕获灰色任务区域内的移动目标,考虑传感器探测概率与虚警概率,提出了重访机制驱动的协同搜索规划（RMD-CSP）方法,以降低目标遗漏与误判概率。考虑无人机飞行性能约束,以最大化任务执行效能为目标建立多无人机协同搜索模型。根据目标先验信息初始化环境搜索信息图（包括目标概率分布图、环境不确定度图与环境搜索状态图）,利用无人机实时探测信息,基于贝叶斯准则持续更新搜索信息图。定制基于环境不确定度更新的重访机制,通过增加长时间未被重访区域的环境不确定度,引导无人机搜索该区域,降低移动目标的遗漏概率;定制基于目标函数权重更新的重访机制,引导无人机快速重访发现新的疑似目标的区域,对疑似目标进行再次确认,减少由于传感器虚警概率造成的目标误判概率。采用滚动时域规划架构,将搜索规划问题分解为一系列短时域规划问题,提升了求解效率。在典型任务想定下,通过数值仿真试验验证了所提方法的有效性。仿真结果表明,RMD-CSP能够在秒级时间内生成每个时域的搜索航迹,相比于光栅式搜索方法与标准的概率启发式搜索方法,能够引导无人机捕获更多的移动目标,同时减少误判次数,有效提升了多无人机协同搜索的任务效能。     

Integrated precision SAR targeting: a SAR targeting simulation

Operational deficiencies exist in the current ability to perform high resolution synthetic aperture radar (SAR) targeting for precision strike missions in difficult electronic countermeasure (ECM) and anti-aircraft environments. Increasingly sophisticated enemy defenses require that fire control information be derived from highly maneuverable aircraft trajectories rather than traditional straight-and-level flight. In addition, target area aspect information is needed to determine optimal attack approaches. These imaging and targeting conditions impose severe performance requirements on the sensor aircraft motion compensation system (MCS), and dictate the need for sophisticated algorithms to counteract ECM threats. This paper describes the development of a user friendly “Design Tool” for modeling and analysis of airborne radar system performance for SAR imaging and Precision SAR Targeting     

一种有限区域运动目标分布模型及计算

潜艇的运动模型对于反潜效能的评估具有重要意义。针对应召搜索时概知潜艇的初始位置、但潜艇运动不确定性的特点,建立了一种有限区域内潜艇的运动和分布模型,并利用数值积分的方法从理论上进行了计算分析,给出了计算的结果。     

模型预测滤波在机载SAR运动补偿POS系统中的应用

 机载合成孔径雷达（SAR）运动补偿用位置姿态系统（POS）的定位精度直接影响SAR成像的效果。为进一步提高POS的导航精度,提出将模型预测滤波（MPF）与扩展卡尔曼滤波（EKF）相结合的方法应用于POS中。该方法不需要假设模型误差为高斯过程,并能够在线实时估计并修正系统模型,有效解决了MPF算法与系统模型不完全兼容的问题。飞行试验结果表明,该方法的收敛速度和滤波精度均明显优于目前工程应用中的KF和EKF,特别是大大提高了POS的定位精度;同时该算法与线性滤波KF的计算量相当,更好地满足了工程应用对导航精度和实时性的要求。     

基于SVD的R-T-S最优平滑在机载SAR运动补偿POS系统中的应用

 机载合成孔径雷达（SAR）运动补偿用位置姿态系统（POS）的导航精度直接影响SAR成像的效果。为进一步提高POS的导航精度和数值稳定性,提出将基于奇异值分解（SVD）的Rauch-Tung-Striebel（R-T-S）最优固定区间平滑应用于POS后处理中。在基于SVD的前向卡尔曼滤波（KF）的基础上,进行了基于SVD的后向R-T-S最优固定区间平滑,获得位置、速度和姿态的最优估计。该方法将原算法中均方误差阵进行奇异值分解,不仅具有很好的数值稳定性和鲁棒性,而且避免了矩阵的求逆。半物理仿真结果表明,该方法在导航精度和数据平滑度上明显优于目前工程中应用的KF,是一种有效的事后处理方法。     

航空器海上飞行空域规划方法

为充分利用海上空域资源,提供海上飞行有效空域活动范围,针对海上空域可用资源问题进行了研究,主要目的是寻找一种有效的海上空域可用资源规划方法。一方面,分析了我国周边海域的海上空域资源,并基于岸基军、民航已规划使用空域的环境数据,结合周边空管雷达监视能力服务范围信息,设计了一种基于图形栅格化法的海上空域资源智能规划方法。另一方面,基于协同区域随机搜索算法,分析计算出适合航空器飞行的海上空域数据,可为航空器海上飞行空域规划提供可靠依据。     

Future X-ray Missions for High Resolution Spectroscopy

The future X-ray missions for high resolution spectroscopy are briefly reviewed. ASTRO-H, planned for launch in 2014, will introduce microcalorimeters for the first time and reveal dynamical motions of hot gas in extended objects. High resolution spectroscopy will also be used for the search of missing baryons with oxygen lines in the local universe. Dedicated X-ray missions are also planned. A very large X-ray observatory IXO, under joint study of NASA, ESA and JAXA, will explore the evolution of the universe using X-ray spectroscopy as a very powerful tool.     

复杂任务场景无人机集群自组织侦察建模与仿真

基于agent的作战建模与仿真是无人机集群侦察效能评估的重要方法,但以往方法中agent行为机制 和模型结构难以描述复杂任务场景下的自组织侦察行为。以反海盗侦察任务为例,提出一种面向集群自组织 侦察的作战建模与仿真方法。设计多agent分层复合行为机制和可组合模块化agent模型结构,通过简单行为 的组合描述复杂行为;通过信息素图模块与航路管理模块的交互,实现自组织航路规划与飞行控制;并通过仿 真算例,来验证方法的可行性,对比分析不同航路规划方法的侦察效能。结果表明:自组织航路规划对中断航 路侦察的动态事件的适应性更强;对于大区域、长时间、低目标密度的区域覆盖侦察任务,仅通过改进航路规划 方法来降低访问间隔难以显著提高目标发现概率。     

Conversations: with Carl Pilcher [interview by Johan Benson

An interview with Carl Pilcher, science program director for solar system exploration at NASA, examines NASA's past, present, and planned missions to explore the solar system. Specific questions relate to the status of current and planned missions, science results of the Pathfinder mission to Mars, cooperation with the Japanese space agency, the status of the search for extraterrestrial life in solar system meteoroids and asteroids, mission size for more in-depth exploration, reports of water on the moon, and the exploration of near-Earth objects.     

基于改进A-Star算法的隐身无人机快速突防航路规划

针对现代战争中隐身无人机（UAV）在高严密的组网雷达防御体系下的生存及突防问题,提出了基于改进A-Star算法的隐身无人机战区突防航路规划技术。首先对隐身无人机突防过程进行了分析建模,分别建立了隐身无人机的运动学模型、动态雷达散射截面特性和组网雷达探测概率的计算模型。然后针对传统算法在解决隐身突防问题时的不足,充分考虑所规划航路时的快速性和安全性要求,设计了改进A-Star算法。在算法中引入了多层变步长搜索策略和无人机的姿态角信息,结合秩K融合准则,通过每段航迹上隐身无人机被组网雷达系统的发现概率来判断新航迹点的可行性。仿真结果表明,改进A-Star算法能够在复杂的组网雷达系统下快速生成更优的战区突防航路,具有一定的应用价值。     

航空发动机传感器解析余度模型的建立方法

为了使用解析余度模型对传感器故障进行诊断,提出了1种基于K-均值聚类与改进微分进化算法优化的极端学习机(IDE-ELM)的发动机传感器解析余度模型建立方法。为避免求解ELM算法时H矩阵奇异,采用K-均值聚类对试验数据进行聚类处理,然后从每类数据中选取1组数据组成训练样本用于训练;利用IDE算法优化ELM的输入层权值和偏置,提高ELM的泛化能力。利用飞行试验数据进行了仿真验证。结果表明:基于K-均值聚类和IDE-ELM设计的传感器解析余度模型具有较高的精度,可用于FADEC系统双通道传感器的故障诊断。     

改进型蚁群算法的全局路径规划仿真研究

针对传统蚁群算法收敛较慢的问题,提出了一种在复杂环境下全局路径规划的改进型蚁群算法。利用链接图法建立了路径规划的空间模型;借鉴狼群分配原则对信息素进行更新;在缩小搜索区域,提高搜索效率的过程中,引入了启发式概率公式和启发函数;通过参数自适应调整策略,进一步对最优解进行了优化。将基于Dijkstra算法的初始路径规划和改进后蚁群算法的规划结果进行了仿真对比,结果表明,改进后蚁群算法的全局优化性能较好,具有一定的有效性和可行性。     

基于LASSO变量选择的航空发动机相似性剩余寿命预测

由于航空发动机监测变量众多,传统方法直接选取性能退化趋势明显的变量进行寿命预测,所以提出一种基于LASSO(least absolute shrinkage and selection operator)的变量选取方法,结合相似性寿命预测方法有效提高了预测精度。基于K-means聚类区分不同工况,对航空发动机多个监测变量根据聚类结果进行变量转换。基于LASSO方法选取最优传感器变量。基于相似性方法进行航空发动机剩余寿命预测。将基于LASSO的变量选取方法与传统的根据退化趋势大小进行选择的方法进行剩余使用寿命预测的结果进行了对比研究。结果表明：基于LASSO选取变量的相似性寿命预测误差的标准差在3种运行周期下分别减少了约1.84、3.46、4.23。     

Smuggling interdiction using an adaptation of the AN/APG-76multimode radar

The AN/APG-76 multimode radar was designed and developed for a multimode attack fighter application requiring rapid search, detection, identification, and precision location of both airborne and surface targets from long standoff ranges under adverse weather conditions. Unique is the radar's ability to generate SAR images of a selected area while simultaneously detecting and tracking all-speed moving objects located within that imaged region. Northrop Grumman Norden Systems has recently upgraded and adapted this radar to smuggling interdiction and related law enforcement missions. These adaptations have added an integrated GPS/INS subsystem for enhanced self-navigation and target location accuracy, a long range wide-band digital data link and ground station for mission control and data dissemination, a 3-D interferometric SAR imaging capability for detailed high resolution topographic mapping, and 1 meter and 0.3 meter resolution SAR modes for positive target identification. Additionally, the radar has been installed into wing-mounted pods and adapted for side-looking and 360 degree coverage applications. Automatic target detection and enhanced-range sea-surveillance and air-targeting modes are also now available through the use of open architecture commercial processors and non-proprietary transportable programming languages     

PDE solution to UAV/UGV trajectory planning problem by spatio-temporal estimation during wildfires

Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicles (UGVs) have been used in research and development community due to their strong potential in high-risk missions. One of the most important civilian implementations of UAV/UGV cooperative path planning is delivering medical or emergency supplies during disasters such as wildfires, the focus of this paper. However, wildfires themselves pose risk to the UAVs/UGVs and their paths should be planned to avert the risk as well as complete the mission. In this paper, wildfire growth is simulated using a coupled Partial Differential Equation (PDE) model, widely used in literature for modeling wildfires, in a grid environment with added process and measurement noise. Using principles of Proper Orthogonal Decomposition (POD), and with an appropriate choice of decomposition modes, a low-dimensional equivalent fire growth model is obtained for the deployment of the space–time Kalman Filtering (KF) paradigm for estimation of wildfires using simulated data. The KF paradigm is then used to estimate and predict the propagation of wildfire based on local data obtained from a camera mounted on the UAV. This information is then used to obtain a safe path for the UGV that needs to travel from an initial location to the final position while the UAV’s path is planned to gather information on wildfire. Path planning of both UAV and UGV is carried out using a PDE based method that allows incorporation of threats due to wildfire and other obstacles in the form of risk function. The results from numerical simulation are presented to validate the proposed estimation and path planning methods.     

非制导炸弹投掷的攻击航迹规划研究

战斗机投掷非制导炸弹,需要预先对攻击航迹做出合理规划,以提高作战效能和生存能力。为此,提出一种战斗机的攻击航迹规划方法。首先建立战斗机运动模型,生成基本战术动作模型库,并将动作组合合成为基本航迹;然后,构建各阶段动作参数的开闭环控制模型,对战术动作进行微观控制,并依据飞行员经验确定参数变化的约束范围;最后,基于非制导炸弹的基本性能和战术运用方式,限定战机的投掷参数,进而完成航迹规划。对机动跃升加俯冲攻击战术进行仿真分析和演示验证。结果表明:所规划航迹满足攻击条件,运动参数符合飞行要求,验证方法可行有效。     

规则回转体自动铺丝轨迹规划与丝束增减

为满足自动铺丝轨迹的满铺覆性要求，针对现阶段自动铺丝轨迹规划存在的不足，提出了不同的丝束增减算法。首先讨论纤维铺放方向的确定和中心轨迹数量的计算，设计不同铺放方向轨迹的生成算法。然后以丝束重叠系数为重要参数，对于纤维局部堆积和空缺问题提出单侧纤维裁剪算法和双侧纤维裁剪算法，并对裁剪后的重叠区域和间隙区域进行面积求解，使得纤维丝束均匀覆于芯模表面。最后基于CATIA CAA二次开发平台，将上述算法集成到纤维铺放CAD系统中，通过运动仿真系统验证算法的正确性。利用提出的丝束增减算法，实现了间隙/重叠区的均匀分布，尽量降低了富树脂区等相关缺陷的聚集对性能的不良影响。     

基于遗传算法的无人机协同侦察航路规划

无人机将成为侦察卫星、有人驾驶侦察机的重要补充与增强手段 ,成为未来战场上广泛应用的一种侦察工具。为了提高无人机 (UAV)的侦察效率 ,在执行侦察任务前必需规划设计出高效的无人机侦察飞行航路。针对这一问题 ,本文提出了一种侦察效率指标评估的计算方法 ,解决了航路规划中的侦察效率量化问题。考虑在大范围任务区域内进行侦察航路优化存在计算的复杂性和收敛性等问题 ,本文采用遗传算法对侦察航路进行了优化处理。通过该方法得到的侦察航路可以有效地提高无人机的侦察效率。     

Convex optimization of asteroid landing trajectories driven by solar radiation pressure

High-area/mass ratio landers driven by Solar Radiation Pressure(SRP) have potential applications for future asteroid landing missions. This paper develops a new convex optimization-based method for planning trajectories driven by SRP. A Minimum Landing Error(MLE) control problem is formulated to enable planning SRP-controlled trajectories with different flight times. It is transformed into Second Order Cone Programming(SOCP) successfully by a series of different convexification technologies. A t...     

基于SAS算法的起飞一发失效应急路径规划方法

为解决起飞一发失效应急程序（EOSID）手动设计的不足,提出一种基于SRTM数据的稀疏A*搜索（SAS）算法的EOSID路径规划方法。首先采用航天飞机雷达地形测绘使命（SRTM）的网格地形数据,结合起飞一发失效相关规章,考虑爬升梯度与保护区限制确定可行搜索空间;然后基于可行搜索空间运用稀疏A*搜索算法搜索应急离场路径,在传统A*算法寻找扩展节点时加入起飞性能约束条件,同时利用地形高程数据进行地形和威胁回避,生成一条三维应急离场航迹;最后利用三次样条曲线对规划的应急离场航迹进行平滑处理。实验结果表明该方法能自动搜索出有效的EOSID三维航迹。