与给定多边形相切的C 3连续B3样条曲线

描述了一种与给定多边形相切的B3样条曲线的算法。在算法中，所有的B3样条曲线的控制点可以通过对多边形的顶点简单计算产生，所构造的曲线对多边形具有保形性，曲线可以局部修改，最后给出了两个算例。     

C~3连续的保凸T-B插值曲线及保形插值算法

基于三角函数的T-B样条曲线虽然具有保凸性,但曲线不通过任何控制点.现在在两个相邻控制点之间插入两个新的控制点,使改造后新的T-B样条曲线插值能够通过原来的控制点,不仅保凸、C~3连续,而且曲线的形状还可作局部修改.给出了T-B样条曲线保形插值算法的保形性条件,并给出了数值计算例子.     

基于分段速度规划的机器人臂运动平滑减振方法

目的 保证机器人臂在生产作业时,其末端速度、加速度、jerk曲线光滑且连续,避免机器人臂在高速运行时发生剧烈振动。方法 提出一种基于分段速度包络轨迹规划方法,该方法与传统样条算法相比,边界条件的确定和计算量大大简化,仅需给定运行轨迹点和期望运行时间,就能实现机器人臂末端轨迹的速度、加速度以及jerk曲线光滑连续,并且通过调节因子和分段时间间隔可以单独控制加速度、jerk曲线的波动幅值和整体运动周期。结果 在双臂并联机器人平台上,依次进行本文算法与三次样条算法的对比实验,以及本文算法在不同调节因子取值下的对比实验。在第1组实验中发现采用本文算法运行机械臂更加平稳,所得轨迹也很平滑,而三次样条算法运行机械臂振动较大;通过第2组对比实验可得,在分段时间间隔一定的情况下,调节因子取值越大机械臂运行就会越流畅,振动也会进一步减小。结论 文中提出的方法能够有效抑制机器人臂振动,可以使机器人高速平稳的运行。     

P-nary细分多层B样条算法及曲面拟合

多层B样条方法在曲面的拟合中提供了新的途径.在证明均匀B样条基函数P-nary细分方程基础上,给出了均匀B样条曲线的P-nary细分控制点之间的计算公式,进而讨论了B样条曲面的P-nary细分问题,并将其用于层次B样条曲面拟合.提出了基于P-nary细分多层B样条曲面拟合算法.该算法加快了层次B样条曲面拟合中网格加细后曲面控制点的计算,数值实例显示所给算法的有效性.     

B-样条曲线升阶的几何收敛性

B-样条曲线的升阶算法是CAD系统相互沟通必不可少的手段之一。B-样条曲线的控制多边形经过不断升阶以后,和Bézier曲线一样都会收敛到初始B-样条曲线。根据双次数B-样条的升阶算法,得到了B-样条曲线升阶的收敛性证明。与以往升阶算法不同的是,双次数B-样条的升阶算法具有割角的性质,这就使B-样条曲线升阶有了鲜明的几何意义。得到的结论可以使B-样条曲线像Bézier曲线一样,通过几何割角法生成。     

基于单位圆弧段逼近的Bézier曲线等距线生成算法

提出一种用四次Bézier曲线逼近单位圆弧段(Unit Circular Arcs)的方法及其详细误差函数分析.使用这种方法,给出一种使用同阶Bézier曲线逼近给定Bézier曲线等距线的算法.在Matlab7.0上实现了该算法,试验表明,新算法比Lee和Ahn所提出的算法有更高的精度和计算效率.由于B样条和NURBS曲线可以认为由多段Bézier曲线组成,因此,新算法为B样条和NURBS曲线等距线的求解提供了一种新的途径.     

平面G2组合三次α-Bézier曲线的几何构造

对三次曲线的几何连续拼接问题做了研究.给出了构造平面G2组合三次α-Bézier曲线的几何算法.这个算法,可以对给定的一组平面控制顶点,方便地构造一条G2三次α-Bézier样条曲线.这种样条保留了B样条、β样条的性质,优点在于保持曲线G2连续,同时通过选取不同的混合因子和形状因子,局部调整曲线的形状,以满足不同的设计要求.     

采用进化算法的空间曲线误差计算

针对非均匀有理B样条(NURBS)表示的空间曲线,利用进化算法对复杂空间曲线形状进行误差计算,阐述了测量点到理论曲线最短距离的求解算法和步骤,构建了误差评定的目标函数.采用实数编码的遗传算法以及微粒群算法对测量点与理论曲线的距离进行计算,并与BFGS方法进行比较.实验结果表明该方法能够快速获得误差评定结果.     

非光滑曲线等长样条逼近

在工程计算中,常常需要对非光滑曲线进行光滑逼近.本文讨论了非光滑曲线在局部小区间上,可用保长度不变的光滑样条曲线来近似的问题.首先利用介值定理,从理论上证明:对定义在有限区间上的任意连续函数,只要曲线长度有限,则在该区间上可用光滑样条曲线来等长近似原曲线.在此基础上,通过选择恰当的插值节点,可以构造出一条唯一的光滑曲线,使其具有和原曲线形同的长度.同时,本文还给出具体构造光滑曲线的方法及详细计算过程.最后的数值结果表明这种方法既简单又行之有效,从而完整地解决了工程计算中曲线近似这一问题.     

基于小波计算的平面物体射影不变性的识别

利用非均匀有理B样条从单幅图像中提取平面物体在射影变换下的几何不变特征,建立经典框架.讨论用小波计算框架的一维和二维特征不变矩的方法以减少计算量.实验结果表明所选取的特征不变量和算法有较小的计算量,对二维平面物体的识别有很好的性能,特别是允许景物中有部分的遮挡物存在.     

Collision-free path planning for coordinate measurement machine probe

This paper focuses on the method of collision-free path of the coordinate measurement machine (CMM). In terms of the collision-free path, the cut face method is used to plan the displaced tracking along the boundary loop. In addition, the 'maximum angle between tangential vectors method' and the heuristic algorithm designated by the displacement characteristic among the feature faces are used to reduce the number of intermediary points. This can also reduce the distance of the displacement. The turning of the probe aims at the contact with the workpiece surface in the normal direction. If the contact position in the normal vector direction is impossible, the probe angle closest to the normal direction is adopted. The normal direction contact with the workpiece can, on the one hand, achieve better measuring position and, on the other hand, accompany with the cut face method and thus make a collision-free path possible.     

Frenet坐标框架下的自适应避障算法

目的 解决包装车间无人运输车辆在沿全局路径行驶过程中,难以同时保证跟随轨迹平滑及规避障碍物的问题。方法 构建Frenet坐标框架,描述车辆位姿与全局路径的相对关系,利用五次多项式生成多条待选局部路径;采用障碍物势场法模糊处理包装车间障碍的外形轮廓,并设置膨胀区间;结合待选路径采样点经过障碍物膨胀区间得到的碰撞值,以及采样点与全局规划路径的偏差值,综合评估路径的避障能力,筛选出最优的避障路径。结果 该算法能根据障碍物的大小,动态解算出与全局路径偏移量小且运动连贯的局部避障路径。结论 所提算法使得车辆完成运输任务时的效果更好,且效率更高。     

Fiberwalk: a random walk approach to fiber representative volume element creation

Attaining accurate representative geometry for a finite element analysis at various scales can be a challenging research task. Added complications arise when the geometry is representing a manufactured or biological composite. In this study, a representative geometry of linear fibrous composites was created at the microscale and is employed to reproduce crystalline microfibril stacking of Kevlar to form a single fibril joined by a non-ordered crystal structure. The structure can have ellipsoidal or rectangular microfibrils stacked in ellipsoidal or rectangular fibrils and may have any desired packing ratio within the 1–100% range. In order to build a fibrous structure along a path, a Random Walk methodology was used. Since the directionality of the fibers is random, but always stepping from one side of the path toward the other, the fibers can wind around each other and tangle or terminate if needed. Another key concept of this method is the addition of a rotation matrix operation for the path of the fibers. This allows the path around the three local coordinates to be in a linear or sinusoidal direction. The resultant geometry produced can represent the tortuous path nanofibrils undergo. Moreover, rotation about the path axis allows for the twisted geometry of ring spun yarn, and metal cable to be reproduced. Inclusion of spherical objects to the path of the fibers has been accommodated to reproduce fiber projection around, or end at, an obstruction in their path. This approach allows representing impurities at the fibril, fiber, or yarn level of composite fabric manufacturing.     

On the importance of path for phase unwrapping in synthetic aperture radar interferometry

Phase unwrapping is a key procedure in interferometric synthetic aperture radar studies, translating ambiguous phase observations to topography, and surface deformation estimates. Some unwrapping algorithms are conducted along specific paths based on different selection criteria. In this study, we analyze six unwrapping paths: line scan, maximum coherence, phase derivative variance, phase derivative variance with branch-cut, second-derivative reliability, and the Fisher distance. The latter is a new path algorithm based on Fisher information theory, which combines the phase derivative with the expected variance to get a more robust path, potentially performing better than others in the case of low image quality. In order to compare only the performance of the paths, the same unwrapping function (phase derivative integral) is used. Results indicate that the Fisher distance algorithm gives better results in most cases.     

Integrated tool positioning and tool path planning for five-axis machining of sculptured surfaces

In this paper a new approach to tool path planning is presented for five-axis machining of sculptured surfaces. The positioning of the cutting tool along a machining pass is determined in an attempt to produce the most efficient machining pass with respect to the entire tool path. In this way the tool positioning strategy is an integral part of the path planning strategy. This differs from current methods, where tool positioning and path planning are two separate tasks. In the present work, various tool orientations are evaluated for cutter locations along the machining pass. The evaluation and eventual selection are made with respect to the completion of the overall tool path. An example part was simulated using the proposed integrated method which resulted in improved efficiency over a more traditional approach. The proposed method was also verified experimentally using cutting tests.     

动态环境下基于自适应步长Informed-RRT*和人工势场法的机器人混合路径规划

为解决移动机器人在动态环境下的路径规划问题,将Informed-RRT*和人工势场法相融合,提出全局与局部规划算法相融合的路径规划方法。首先,针对Informed-RRT*算法采样效率低,以及得到路径不满足机器人运动学约束的问题,采用目标偏置法与自适应步长法,减少冗余搜索与不必要树的生长;同时,引入走廊优化与时间重分配法,优化路径节点,使路径更加平滑。其次,针对人工势场法易陷入局部极小值和目标点附近不可达的问题,采用平滑窗格策略,增设全局路径子目标点,使机器人能够逃离局部极小值,完成规划任务。仿真结果表明,静态环境中自适应步长Informed-RRT*算法相比于Informed-RRT*算法求解时间缩短了71.98%;动态环境中,混合算法相比于人工势场法,搜索时间缩短了15.4%,路径长度缩短了11.1%。     

一种稳健的运动探测与背景维护算

针对家庭安全监控需要，提出一种稳健的运动对象探测算法。首先给出三帧差运动对象检测算法，同时利用像素的亮度分量和色度分量避免由于大尺寸对象闯入引起图像背景像素变化造成运动对象检测的失败。然后提出一种三关键背景维护算法，用于清除背景振动像素，减少虚假报警。实验表明，这种运动探测算法对视频监控系统消除虚假异常报警具有良好的通用性，实时性和稳健性。     

Path planning of functionally graded material objects for layered manufacturing

Path planning is a fundamental process-planning task in layered manufacturing. However, currently, most researches focus on path planning of homogeneous objects and few approaches for path planning of functionally graded material objects have been reported in the literature. Unlike homogeneous objects, functionally graded materials objects contain continuously varying material information as well as geometry information, which makes their path planning more complex and difficult than those of homogeneous objects. This paper presents an approach for path planning of functionally graded material objects. First, the continuous material distribution in each layer is changed into discrete step-wise gradings by subdividing the functionally graded materials slice into multimaterial subregions, which can be regarded as homogeneous material regions so that the functionally graded material layer can be built by mixing materials into desired volume ratio. Then the tool paths and process parameters for each multimaterial subregion in the slice are generated separately. An algorithm that summarized the procedure is described and an example is also presented.     

Color rendering properties of interior lighting influenced by a switchable window

Switchable windows are being installed into modern buildings. We report on how the variable transmittance of gasochromic and electrochromic switchable windows changes the color rendering properties of the daylight passing through these windows. We present a series of color rendering examples. Correlated color temperature and color rendering indices are insufficient to describe the color properties of this filtered light, as these indices are beyond the applicability limits. We obtain more reliable results using the color shifts (deltaE) of different objects. We find that as the transmittance of the window changes, each surface color moves along a path of the CIELAB space in the same direction for both switchable units. The direction and distance moved differs among test-color samples. The appearance of an array of colored objects is strongly distorted at higher coloration states of the windows.     

Invariant pattern recognition based on centroids

A new method for pattern recognition that is invariant under changes of position, orientation, intensity, and scale is presented. The centroids of objects provide unique points that are related to the energy distribution. For obtaining more such unique points a conformal transform can be used to rearrange the energy distribution of the object. By means of the conformal transform many different centroids can be produced from the same object. A useful pattern-recognition and object-registration method that yields a position-, rotation-, intensity-, and scale-invariant feature vector based on these centroids can be created.