单目视觉里程计/惯性组合导航算法

提出一种单目视觉里程计/捷联惯性组合导航定位算法.与视觉里程计估计相机姿态不同,惯导系统连续提供相机拍摄时刻对应的三维姿态,克服了单纯由视觉估计相机姿态精度低造成的长距离导航误差大的问题.通过配准和时间同步,用惯导系统解算的速度和视觉里程计计算的速度之差作为组合导航的观测量,采用Kalman滤波修正组合导航系统的误差,同时估计视觉里程计标度因数误差.分别在室内外不同环境下进行了22 m的推车实验和1412m的跑车实验,定位误差分别为3.2%和4.0%.与Clark采用姿态传感器定期更新相机姿态估计结果的方法相比,单目视觉里程计/惯性组合导航定位精度更高,定位误差随距离增长率低,适合步行机器人或轮式移动机器人在复杂地形环境下车轮严重打滑时的自主定位导航.     

基于非线性优化的单目视觉/惯性组合导航算法

针对自主移动机器人在GPS拒止环境下准确实时定位的问题,提出一种单目视觉/惯性组合导航算法。为解决视觉/惯导工作频率不一致问题,利用预积分技术预先处理惯性测量值;引入了一种快速高精度线性初始化方法,分步估计初始系统尺度、重力方向、速度及零偏;结合全局地图三维点约束削弱累积误差,基于是否更新地图点,构建了两种基于非线性优化的单目视觉/惯性紧耦合模型,以保证导航的局部精度及全局一致性。实验结果表明:初始化方法可快速(15 s内)实现高精度状态初始化;与单目视觉导航算法相比,提出的算法不仅可获取绝对尺度信息,且导航精度更高;与传统滑动窗口非线性优化方法相比,提出的算法在窗口优化过程中加入全局地图三维点约束,可有效削弱累积误差,验证了算法的正确性和可行性。     

一种单目视觉ORB-SLAM/INS组合导航方法

针对惯性/卫星组合导航系统在卫星导航失效时无法使用的问题,提出了单目视觉ORB-SLAM/INS组合导航方法,用于扩展组合导航系统在强干扰环境和室内环境的应用范围。该算法分为两个阶段:初始化阶段,当ORB-SLAM形成闭环时设计算法在线估计单目视觉ORB-SLAM算法的尺度因子;导航阶段,ORB-SLAM系统输出的位置信息经过尺度变换后作为观测量进行卡尔曼滤波,估计INS导航系统的误差状态量从而修正惯导系统的误差。设计了硬件和软件平台对提出的组合导航方法进行试验验证。跑车实验结果表明:所设计的ORB-SLAM/INS组合导航系统具有较高的定位精度,导航时间6 min定位误差为1.162 m,且不随时间漂移,具有很强的应用价值。     

基于先验特征地图匹配约束的视觉/惯性定位方法

单目视觉/惯性定位系统由于其低成本和高自主性被广泛使用.然而视觉/惯性里程计在长时间的运行后,会产生较大的累积误差,无法提供准确的位姿估计,并且在光线、路况复杂的环境系统存在特征点跟踪失败,无法解算当前位置,稳定性不高的问题.针对以上问题,提出了基于非线性优化的先验特征地图匹配方法,利用单目视觉/惯性SLAM建立的位姿...     

基于交互式多模型的车载INS/OD/GPS的定位方法

INS/GPS组合导航时,GPS信号会由于环境因素或是其他信号干扰导致信号丢失、信号跳变等故障。针对INS/GPS组合导航中,在GPS出现离群故障难以得到满意定位性能的问题,提出了一种基于交互式多模型滤波的车载INS/OD/GPS定位方法。引入里程计对INS/GPS组合导航进行辅助,基于INS/GPS、INS/OD两种组合导航模型,利用交互式多模型算法交互更新模型概率,通过模型概率的差异判断主滤波器INS/GPS组合过程中的GPS量测故障,并在滤波过程中对故障量测加以修正,以此来实现车辆定位系统的良好定位性能。实验结果表明,在里程计量测辅助条件下,当GPS出现离群点时,系统平均定位精度比INS/GPS单一滤波器条件下的定位精度提升60%以上。     

基于多传感器信息融合的双足机器人自主定位

针对单目视觉SLAM应用于双足机器人过程中易出现跟踪失败和传统位姿传感器累积误差较大的问题,提出了一种融合三种传感器信息的机器人定位方法。首先获取机器人关节编码器数据,运用运动学求解输出位置信息,再由基于特征识别的SLAM系统估计机器人头部单目相机位姿,经过尺度恢复和坐标系校正后输出位置信息,同时获取惯性测量单元的偏航测量,最后通过扩展卡尔曼滤波算法融合三种信息,从而更加精确稳定地对机器人进行定位。以双足机器人NAO为研究平台,在室内进行多组实验对所提出的方法进行验证,结果表明融合所得结果与传统位姿传感器输出信息相比,均方根误差明显降低;相比单目视觉SLAM跟踪失效率至少降低88%。     

基于双向重投影的双目视觉里程计

针对视觉里程计中的特征点漂移与累积误差问题,提出一种基于双向重投影的双目视觉里程计方法。为降低特征点匹配的高错误率及改善匹配效率,提出一种改进的随机采样一致(RANSAC)算法。该算法依据特征点寿命长短、环形匹配、优质匹配等方法选取优质特征点以降低特征点样本容量,从而改善RANSAC算法的效率;为抑制视觉里程计累积误差、提升位姿估计精度,提出一种双向重投影的位姿估计算法,该方法通过解算下一帧图像中的最小化重投影误差得到相机位姿以更新空间点集,更新后的空间点重新投影至当前帧图像以提升位姿解算精度。最后利用KITTI数据集进行了仿真。结果表明该算法的150 m内定位误差在X方向小于1 m,在Z方向小于0.9 m;相对于ORB-SLAM2算法,定位精度提升了约20%,系统的稳定性更好,有效削弱了累积误差,验证了该算法的正确性与可行性。     

一种微型无人机单目视觉SVO/INS组合导航方法

针对未知环境中微型无人机自主导航问题,提出一种单目视觉SVO/INS组合导航算法。使用半直接视觉里程计(SVO)作为视觉里程计前端,首先对SVO的初始化模块、关键帧选择及跟踪失败处理策略进行了重新设计,实现了SVO在摄像头前视场景中的应用。之后,设计了误差状态卡尔曼滤波器,基于惯性测量单元的测量值进行状态预测,采用SVO的估计结果作为观测值进行状态更新。使用EuRoc数据集对所提算法进行验证,结果表明所提算法可实现对无人机的位置、姿态和速度以及未知尺度、惯性测量单元零偏、重力方向的估计。导航时间3 min位置估计误差为0.15 m,俯仰和横滚角估计误差小于0.5°,航向角估计误差小于1.2°。所提方法具有计算量小、实时性好的优点,适用于机载计算资源有限的微型无人机导航。     

基于车轮安装惯性测量单元的车载组合导航

微电子机械系统惯性测量单元(MEMS-IMU)因其精度低且温度敏感性大,以及传统安装方式通常存在航向角不可观等原因,在车载导航和轮式机器人应用场景中表现为定位误差迅速累积。为解决上述问题,设计了一种将MEMS-IMU安装在载体车轮中心的导航方案。利用MEMS-IMU与车轮一起旋转的运动特性,提取出运动速度信息,并结合运动约束构造了三维速度观测信息,从而改善最终的组合导航性能。该方案从安装方式和算法设计层面取代了对里程计传感器的依赖。基于轮式机器人的实测结果表明,该方案的平面定位误差(均方根值)相比于传统安装MEMS-IMU的导航方案降低了58.53%,相比于传统MEMS-IMU和里程计组合导航方案降低了29.67%。     

基于运动学非完整约束的里程计参数在线辨识

针对里程计标度因数误差以及安装误差对捷联惯导/里程计组合导航精度存在较大影响的问题,提出一种基于运动学非完整约束的里程计参数在线辨识方法。通过建立航位推算误差模型,利用里程计输出,计算车辆的向心加速度,并与捷联惯导的加速度计输出计算的向心加速度做差作为量测之一;将里程计与捷联惯导输出的速度做差作为量测之二;通过卡尔曼滤波实现里程计标度因数误差以及安装误差的在线辨识。仿真结果表明,该方法组合导航相比速度组合导航,当车辆行驶960s,其东向定位精度提高19.6 m,北向定位精度提高14.2 m。仿真和车载试验结果均表明该方法能实现里程计参数的有效辨识以及航向安装偏差角的估计,有助于提高组合导航精度。     

Group properties of utt=[F(u)ux]x

The group properties and the associated Lie algebra are developed for the subject quasilinear wave equation, for arbitrary f[fεC²(R), f > 0, f ≠ 0]. From the resulting information sets of explicit invariant solutions are constructed for wave propagation in gases and for the transonic equation.     

Exact solution of the non-linear differential equation RR̈ + 32R2 − AR−4 + B = 0

The non-linear equation $\text{R}\text{R}\text{?}\text{+}\text{3}\text{2}\text{R}{}^2\text{- AR}{}^{\mathrm{?4}}\text{+ B = 0}$ is shown to represent simply periodic motion with a minimum at R₁ and a maximum at R₁R₀ or a maximum at R₁ and a minimum at R₁R₀^?1. R₀ is a function of the ratio $\text{A}\text{B}$ and is greater than 1 for $\text{A}\text{B}$ > 1 and less than 1 for $\text{A}\text{B}$ > 1. The period of the motion satisfies the simple relation T(R₀^?1) = R₀^?1T(R₀). The exact solution to the above equation is represented in terms of elliptic integrals of the first and second kinds and a simple algebraic function.     

Explicit expressions of S(v), H(v) and L(v) for anisotropic elastic materials

The three matrices L(v), S(v) and H(v), appearing frequently in the investigations of the two-dimensional steady state motions of elastic solids, are expressed explicitly in terms of the elastic stiffness for general anisotropic materials. The special cases of monoclinic materials with a plane of symmetry at x₃ = 0, x₁ = 0, and x₂ = 0 are all deduced. Results for orthotropic materials appearing in the literature may be recovered from the present explicit expressions.     

惯导／GNSS／罗兰C／航姿系统组合导航方案

为保证飞行器的导航精度及可靠性,采用了惯导/GNSS/罗兰C/航姿系统的组合导航。该组合导航采用开环组合方式,将多系统导航信息在计算机中进行综合,提高导航信息精度。仿真试验结果说明,所设计的组合导航系统是可行的,能满足任务要求。     

Phase portraits and bifurcations of the non-linear oscillator: ẍ + (α + γx2 + βx + δx3 = 0

The non-linear oscillator ? + αx + γx²x + βx + δx³ = 0 is studied using the methods of differentiable dynamics to obtain qualitative behaviour. The case x, β<0; γ, δ> 0 is considered in some detail; it has physical relevance as a simple model in certain now-induced structural vibration problems in which the structural non-linearities act to maintain overall stability. The presence of local and global bifurcations is detected and their physical significance discussed.     

Rheology/morphology/flow conditions relationships for polymethylmethacrylate/rubber blend

Viscoelastic behavior, phase morphology and flow conditions relationships in polymer/rubber blends have been investigated. The importance of such correlations is illustrated on polymethylmethacrylate (PMMA)/rubber blends subjected to different flow conditions both under small and large deformations. In small-amplitude oscillatory shear (the morphology does not change during the flow) the elastic modulus G of the concentrated blends shows a secondary plateau, G _p , in the low frequency region. This solid-like behavior appears for rubber particle contents beyond the percolation threshold concentration (15%). Morphological observations revealed that for concentrations higher than 15%, the particles are dispersed in a three-dimensional network-type structure.In capillary flow it was found that the network-type structure was destroyed and replaced by an alignment of particles in the flow direction. This morphological modification resulted in a decrease in both viscosity and post-extrusion swell of the blends. Morphological observations revealed that the ordered structure in the flow direction was concentrated only in the skin region of the extrudate, where the shear stress is higher than the secondary plateau, G _p . A simple kinetic mechanism is proposed to explain the observed morphology.Similarly, steady shear measurements performed in the cone-and-plate geometry revealed alignment of particles in the flow direction for shear stress values higher than G_p.Presented in part at the Symposium Recent Developments in Structured Continua Montréal (Canada) 26–28 May 1993 and at the 45th Canadian Chemical Engineering Conference, Quebec, October 15–18 (1995)     

溅射沉积WSx/Ni/a-C/Ni多层膜的微结构及大气环境摩擦学性能

采用磁控溅射技术在硅基体上交替沉积WS₂、Ni及非晶碳(a-C)层制备出不同Ni层厚度的WS_x/Ni/a-C/Ni多层膜. 利用扫描电子显微镜、X射线衍射仪、拉曼光谱仪、X射线光电子能谱仪等研究了多层膜的成分及微观结构;通过纳米压痕仪、划痕仪和球盘式摩擦磨损试验机等评价了薄膜的力学及大气环境中的摩擦学性能. 结果表明:随着Ni层厚度的增加,WS_x/Ni/a-C/Ni多层膜的致密度下降,n_s/n_w值(S与W原子百分数之比)由0.84降至0.73,WS_x层以微晶或非晶的形式存在. 多层膜的硬度和磨损率均随Ni层厚度的增加先降低后升高,但摩擦系数由0.22升至0.38,结合力先增大后减小. Ni层厚度为6 nm的多层膜的硬度可达13.4 GPa,抗氧化性能和耐磨性最优,磨损率仅为9.47×10–14 m3/(Nm).      

机翼-机身-挂架-短舱复杂流场数值模拟

本文采用分区搭接网格技术,对机翼/机身/挂架/短舱复杂组合体进行网格分布,通过分析计算网格对结果的影响,探讨了网格的划分.基于Roe的近似黎曼解的方法,采用S-A湍流模型,通过求解N-S方程,对该组合体外流场/发动机短舱内流场进行一体化数值模拟,与相应风洞实验数据进行了比较与分析,取得了与实验数据较为吻合的结果.与无发动机短舱的组合体的气动特性进行比较,分析了短舱对翼身组合体的气动干扰.     

On free energy-based formulations for kinematic hardening and the decomposition F = fpfe

Within the framework of linear plasticity, based on additive decomposition of the linear strain tensor, kinematical hardening can be described by means of extended potentials. The method is elegant and avoids the need for evolution equations. The extension of small strain formulations to the finite strain case, which is based on the multiplicative decomposition of the deformation gradient into elastic and inelastic parts, proved not straight forward. Specifically, the symmetry of the resulting back stress remained elusive. In this paper, a free energy-based formulation incorporating the effect of kinematic hardening is proposed. The formulation is able to reproduce symmetric expressions for the back stress while incorporating the multiplicative decomposition of the deformation gradient. Kinematic hardening is combined with isotropic hardening where an associative flow rule and von Mises yield criterion are applied. It is shown that the symmetry of the back stress is strongly related to its treatment as a truly spatial tensor, where contraction operations are to be conducted using the current metric. The latter depends naturally on the deformation gradient itself. Various numerical examples are presented.