Particle filter-based real-time phase line bias estimation for GNSS-based attitude determination with common-clock receivers

Global navigation satellite system (GNSS)-based attitude determination has been widely adopted in a wide variety of terrestrial, sea, air, and space applications. Recently, the emergence of commercial multi-GNSS common-clock receivers has brought new opportunities for high-precision GNSS-based attitude determination with single-differenced (SD) model. However, the key requirement of using this approach is the accurate estimation of the troublesome line bias (LB) in real-time. In this contribution, we propose a particle filter-based real-time phase LB estimation approach that apply to SD model with single-system single-frequency observations from common-clock receiver. We first analyzed the relationship between the integer ambiguity ratio value and the phase LB. It is proved that the accuracy of a given phase LB value can be qualified by the related ambiguity resolution ratio value, and the normalized ratio value can therefore be used to represent the likelihood function of observations. Then, we presented the particle filter-based real-time phase LB estimation procedure, and assessed its performance using GPS L1/BDS B1I observations from two datasets collected with different types of common-clock receivers in terms of the accuracy and convergence time of phase LB estimation, the computation load, and the positioning and attitude determination accuracy with respect to the double-differenced (DD) model. Experimental results demonstrated that the phase LB could be accurately estimated with short convergence time (generally within 15 epochs). Moreover, compared with the classical DD approach, the particle filter-based SD approach delivers comparable positioning root-mean-square (RMS) errors in the North and East components but significantly smaller RMS errors in the Up component. Accordingly, the achievable yaw accuracy is comparable whereas the pitch accuracy is remarkably improved. The improvements of positioning accuracy in the Up component and pitch accuracy are approximately 35.7 % to 63.7 %, and 33.3 % to 63.1 %, respectively. Additionally, the single-epoch computation time with our particle filter-based SD approach is generally 0.08 s, which is obviously larger than the DD approach but could still meet the requirements of real-time applications below 10 Hz sampling.     

采用局部约束线性编码的像素级舰船目标图像融合

为了提升视觉信息在低能见度条件下的适用性,提出了一种采用局部约束线性编码的像素级舰船目标图像融合方法。首先,采用K均值奇异值分解算法从海量的训练样本中完成过完备字典的学习。其次,在考虑着舰导引实时性任务需求的前提下,采用局部约束线性编码完成融合系数的非迭代求解,相较于压缩感知理论架构下的匹配追踪算法,极大地降低了计算复杂度。此外,设计了一种基于融合系数最大绝对值的融合规则,并根据过完备字典和融合后的局部约束线性系数实现融合图像的重建。最后,利用圆周滤波器提取舰船图像的候选区域。大量试验结果表明,在融合图像质量方面,所提方法在保留高频细节的同时将图像中的舰船目标有效增强了,MI、Q_w、Q^AB/F等客观评价指标优于同类算法;在实时性方面,所提算法的计算速度相比于采用匹配追踪算法的图像融合方案有明显提升。     

基于网络架构的箭载图像采集系统设计及硬件实现

图像采集系统负责监控和记录运载器上多个重要点位的视频信息。文中介绍了一种基于网络架构的箭载通用图像采集系统,负责采集高清图像并无损存储在图像压缩器内,将高清图像压缩后经遥测传输至地面,由于带宽有限,可实时切换下传的图像信息,同时支持通过网络实时访问箭上图像信息和事后图像数据回读功能,为评估箭上的工作状态或分析故障提供了保障。     

Numerical study on combustion efficiency of aluminum particles in solid rocket motor

The combustion of aluminum particles in solid rocket motor plays an important role in energy release of propellants. However, due to the limited residence time, aluminum particles may not be burned completely, thus hindering the improvement of specific impulse. This study aims to explore the characteristics of aluminum combustion efficiency and its influencing factors by experiments and numerical simulations, providing a guideline for engine performance improvement. As an input of simulation, the initial agglomerate size was measured by a high pressure system. Meanwhile, the size distribution of the particles in plume was measured by ground firing test to validate the numerical model. Then, a two-phase flow model coupling combustion of micro aluminum particle was developed, by which the detailed effects of particle size, detaching position and nozzle convergent section structure on aluminum combustion efficiency were explored. The results suggest that the average combustion temperature in the chamber drops with increasing initial particle size, while the maximum temperature increases slightly. In the tested motors, the aluminum particle burns completely as its diameter is smaller than 50 μm, and beyond 50 μm the combustion efficiency decreases obviously with the increase of initial size. As the diameter approaches to 75 μm, the combustion efficiency becomes more sensitive to particle size. The combustion efficiency of aluminum particle escaping from end-burning surfaces is significantly higher than that from internal burning surface, where the particle combustion efficiency decreases during approaching the convergent section. Furthermore, the combustion efficiency decreases slightly with increasing nozzle convergent section angle. And theoretically it is feasible to improve combustion efficiency of aluminum particles by designing the convergent profile of nozzle.     

Pose matters: Pose guided graph attention network for person re-identification

Person re-Identification (reID), aiming at retrieving a person across different cameras, has been playing a more and more important role in the construction of smart city and social security. For deep-learning-based reID methods, it has been proved that using local feature together with global feature could help to give robust representation for person retrieval. Human pose information can provide the locations of human skeleton to effectively guide the network to pay more attention to these key areas, and can also help to reduce the noise distractions from background or occlusions. Based on human pose, a Pose Guided Graph Attention (PGGA) network is proposed in this paper, which is a multi-branch architecture consisting of one branch for global feature and two branches for local key-point features. A graph attention convolution layer is carefully designed to re-assign the contribution weight of each extracted local feature by modeling the similarity relations. The experimental results demonstrate the effectiveness of our approach on discriminative feature learning. Our model achieves the state-of-the-art performance on several mainstream evaluation datasets. A plenty of ablation studies and different kinds of comparison experiments are conducted to prove the effectiveness of this work, including the tests on occluded datasets and cross-domain datasets. Moreover, we further design supplementary tests in practical scenario to indicate the advantage of our work in real-word applications.     

基于形态学算法的2219铝合金钨极氦弧焊熔池图像特征提取

根据2219铝合金氦弧焊特点,构建了焊接过程实时视觉传感系统,获取了熔池清晰图像;为获得2219铝合金氦弧焊熔池尺寸参数,实现对熔池特征的分析,采用了形态学算法,提取了典型熔池图像边缘形状,并获得了熔池宽度特征尺寸;通过试验验证,提取的熔池图像宽度特征能够准确反映出熔池尺寸信息及变化趋势。研究结果表明:形态学算法准确度高、计算速度快,满足2219铝合金氦弧焊熔池图像实时处理需求,不同阈值及结构元素对图像特征提取效果均会产生较大影响,研究阈值自适应算法,可以满足工程上焊缝熔池特征实时视觉监测的复杂需求。     

双补燃室固冲凝聚相颗粒分布对比试验研究

针对固冲发动机燃烧效率较低以及凝相粒子分布规律不清楚的现状，本文建立了一套评价固冲发动机燃烧效率的双补燃室试验装置和研究凝相颗粒分布的取样装置。对比研究了结构相同的补燃室内的参数，其误差小于0.5%，说明该装置可以用于精确对比不同补燃室的燃烧效率。然后应用该装置完成了补燃室长度分别为600mm、800mm、1000mm，进气角度分别为30°、45°、60°的对比试验，分析了粒子的粒度分布规律和化学成分变化规律，该规律与数值分析的规律相同。试验结果表明：在进气角度不变情况下，补燃室长度增加有利于提高燃烧效率，燃烧室压强提高，凝相粒子直径明显减小，B粒子燃烧更充分；在补燃室长度为1000mm的情况下，进气角度增加有利于燃烧效率增加，燃烧室压强提高，粒子直径减小，B粒子燃烧更充分。双补燃室发动机是评价补燃室燃烧效率的有效装置，粒度分布和燃烧产物的化学成分有助于补燃室结构设计。     

改进的Chan-粒子滤波算法超宽带室内三维定位

针对超宽带室内定位时容易受到非视距误差的影响，导致定位精度大大下降，甚至无法准确定位的问题，提出了一种融合粒子滤波的改进Chan定位新方法。首先，在Chan定位方法中引入模拟退火算法，对到达时间差（TDOA）算法的测量值进行优化，获得初步定位的最优解；然后，采用粒子滤波对测量值进行再次优化，粒子重采样后得到的中心位置即为目标节点的精确位置。仿真和实验结果均表明，该算法能有效提高在非视距环境下超宽带室内定位的精度，较好地消除非视距误差的干扰。