基于2003年标准数据的中国飞行员人体模型

首先介绍了CATIA 软件中的人体模型及其功能; 然后对软件中的加拿大、法国、韩国、日本人体数据进行了分析; 在此基础上就软件中建立新人体模型需要的人体数据与2003 年测量得到的中国飞行员的人体数据做了比较; 最后对中国飞行员人体数据做了选择和补充, 同时在CATIA 平台上建立了中国飞行员人体模型。该模型可以满足主机厂所在初步总体阶段的工效分析、评价工作。     

基于惯性动作捕捉的人体运动姿态模拟

针对人体运动姿态编辑的自由性,提出一种人体运动姿态模拟方法。该方法采用贝塞尔曲线和数值数据编辑人体运动姿态。根据人体运动的特点,在VC++中运用OpenGL构建虚拟人体模型,利用动作捕捉技术设计人体运动姿态的模拟程序。结合人体关节正常活动范围,对主要关节点的运动姿态进行分析,结果表明,该方法能有效利用人体运动数据,驱动虚拟人体模型。     

飞行员数字化三维人体模型研究

根据直升机仿真系统的特点,简化飞行员人体机械模型,列出参数化飞行员人体模型有限点测量数据变量,在CATIA V5软件平台上建立了数字化人体曲面几何模型.在此基础上建立运动学模型,可以应用于直升机人机工效设计和训练模拟等方面.     

基于微型传感器驱动的三维实时运动人体模型

建立一种运用于人体运动捕获与再现系统的人体模型,实现人体运动实时控制模型的运动.根据微型传感器人体运动捕获与再现系统的特点,提出一种三维人体模型的层次化结构建模方法和多边形分组皮肤变形方法.该方法以生物动力学为基础,根据人体关节的运动特征进行分层表示,采用传感器数据驱动并控制人体骨架模型的运动以及皮肤模型的变形,实时再现真实人体运动.实验表明,新建模方法适用于传感器数据驱动人体模型的实时运动再现,并能够逼真地体现人体运动的生物力学特征.     

Kinect动作捕捉到Gazebo人体模型数据转换的研究与开发

人机协作中人与机器人之间没有物理防护栅,存在碰撞风险,研究通过Kinect采集人体运动,在仿真软件Gazebo中分析工作空间中人与机器人之间的关系来防止碰撞。由于Kinect人体模型与Gazebo人体模型存在差异,将Kinect采集的人体运动数据以四元数形式存储在CSV文件中,通过编制转换软件,从而能够正确驱动Gazebo中人体模型运动,为将来实现人机协作的仿真研究奠定基础。     

基于国标人体测量学数据的虚拟人缩放方法

具有准确人体测量学数据的虚拟人模型在虚拟制造和虚拟维修领域有重要的理论和现实意义。此外,根据某些人体测量学数据,参照典型的人体模型生成特定的虚拟人,即实现人体模型的参数化控制,对于合理地进行人机功效分析也有重要的应用价值。该文在分析我国成年人体测量学数据标准的基础上,通过研究Jack软件中人体模型表示方法和人体测量学数据之间的映射关系,提出了一种基于人体测量学数据的人体模型动态缩放方法。利用该方法在Jack软件中建立了基于国标人体测量学数据的人体模型动态缩放子系统。     

人体运动数据重构方法进展*

运动捕捉系统产生的人体运动数据是标记点在运动序列中的位置数据,用于驱动人体模型产生真实感的动画。在对近几年有关人体运动数据重构的文献进行综合和分析的基础上,首先对人体运动数据重构进行了问题描述,并对人体运动数据在重构过程中难以避免的噪声问题和特征点(虚拟空间中的标记点)缺失问题的研究分别进行了总结和分析;然后对人体运动数据获取的光学式原型捕捉系统开发的研究进行了讨论,评述了人体运动数据驱动人体几何模型的相关研究;最后对未来研究提出了一些展望。     

基于单台Kinect的高精度个性化人体建模方法

提出一种基于单台Kinect的高精度三 维人体建模方法。首先使用一台Kinect对人体头部进行扫描,以获取高精度的头部点云数据 ;其次,在保持头部精确度的基础上对所获取的点云数据进行采样预处理;最后利用层次化 的紧支撑径向基函数(CS-RBFs)将采样后的点云与已有的人体躯干进行拟合获得3D人体模型 。仿真结果表明,该方法能够提高人体模型的精确度并提高建模速度。     

融入变分自编码网络的文本生成三维运动人体

目的 针对现有动态三维数字人体模型生成时不能改变体型、运动固定单一等问题,提出一种融合变分自编码器（variational auto-encoder,VAE）网络、对比语言—图像预训练（contrastive language-image pretraining,CLIP）网络与门控循环单元（gate recurrent unit,GRU）网络生成运动三维人体模型的方法。该方法可根据文本描述生成相应体型和动作的三维人体模型。方法 首先,使用VAE编码网络生成潜在编码,结合CLIP网络零样本生成体型与文本表述相符的人体模型,以解决蒙皮多人线性（skinned multi-person linear,SMPL）模型参数不合理而生成不符合正常体型特征的人体模型问题;其次,采用VAE网络与GRU网络生成与文本表述相符的变长时间三维人体姿势序列,以解决现有运动生成方法仅生成事先指定的姿势序列、无法生成运动时间不同的姿势序列问题;最后,将体型特征与运动特征结合,得到三维运动人体模型。结果 在HumanML3D数据集上进行人体生成实验,并与其他3种方法进行比较,相比于现有最好方法,R精度的Top1、Top2和Top3分别提高了0.031、0.034和0.028,弗雷歇初始距离（Fréchet inception distance,FID）提高了0.094,多样性提高了0.065。消融实验验证了模型的有效性,结果表明本文方法对人体模型生成效果有提升。结论 本文方法可通过文本描述生成运动三维人体模型,模型的体型和动作更符合输入文本的描述。     

服装CAD中个性化三维人体建模

针对三维服装仿真中对各种不同体态特征的人体模型的需求,给出了一种个性化三维人体建模方法。首先对一系列具有不同特征尺寸的成年女性的人体扫描数据进行简化处理,建立具有一致拓扑的人体模型;然后根据不同人体的对应数据点及其相应的特征尺寸,生成各个简化数据点随特征尺寸变化的规律。利用此变化特性,对参考人体模型进行变形得到新尺寸下的人体模型。该方法已在所开发的三维虚拟试衣系统实现,并取得了较为理想的试验效果。     

A field study on braking responses during driving. I. Triggering and modulation

The present experiment was carried out to explore the response of driving subjects to emergency braking. The field trial consisted of driving behind a leading vehicle while the following drivers' responses were recorded by telemetry. A group of 51 individuals performed a series of trials at two driving speeds (60 and 80km/h), two following distances (6 and 12 m), and two braking conditions (real and dummy braking). Not all of these subjects completed all conditions or the minimum number of trials. The dependent variables were the total braking time (TBT) and its subcomponents: braking reaction time (BRT), and accelerator-to-brake movement time (MT). These data were analysed in three separate three-way ANOVAs with repeated measures on all factors. The results showed that when subjects were not aware of the forthcoming braking, the distance and braking conditions had major effects on all dependent variables. At the shorter following distance drivers reacted and moved faster. Similarly, when the brakes were real compared with dummy (i.e. brake lights only) drivers reacted faster. In addition, drivers reacted to onset of brake lights in 83% of the cases when dummy braking was applied, compared with 97% when real brakes were applied. Speed of driving did not show any significant effects and did not appear to influence the cognitive or attentional set to anticipate an emergency manoeuvre. These findings suggest that changes in angular velocity during optic expansion of the leading vehicle may be used as a cue to modulate braking movement, while onset of brake lights alone may be enough to trigger a ‘ballistic’ preventive response.     

未知相对度非线性系统的一阶D型迭代学习控制

The classical D-type iterative learning control law depends crucially on the relative degree of the controlled system, high order differential iterative learning law must be taken for systems with high order relative degree. It is very difficult to ascertain the relative degree of the controlled system for uncertain nonlinear systems. A first-order D-type iterative learning control design method is presented for a class of nonlinear systems with unknown relative degree based on dummy model in this paper. A dummy model with relative degree 1 is constructed for a class of nonlinear systems with unknown relative degree. A first-order D-type iterative learning control law is designed based on the dummy model, so that the dummy model can track the desired trajectory perfectly, and the controlled system can track the desired trajectory within a certain error. The simulation example demonstrates the feasibility and effectiveness of the presented method.     

回归模型中哑变量的相对重要性指数

为在回归模型中描述定性属性,通常需要引入哑变量。对含哑变量的回归方程,提出描述不同哑变量在回归方程中不同重要程度的方法。该方法分解出含哑变量的回归方程中哑变量部分和非哑变量部分的回归平方和,计算这两部分在该回归方程中所起作用的占比,将该占比设计为各哑变量在回归方程中的相对重要程度指数。在近10万笔的Lending Club和Prosper网络借贷数据集上,所进行的挖掘借款用途对借款成功率、信用等级对借款利率的影响程度的实验结果表明,与传统回归方程仅提供哑变量前的系数却不能展现其重要程度相比,所提方法展现出不同哑变量的不同重要程度,为定量分析回归方程中定性自变量对因变量的影响程度提供了重要的手段。     

全局分层关联网络流在多目标跟踪中的应用

整体上将多目标跟踪问题转化为图的问题。首先采用经典的分层思想，建立两层跟踪框架，并将目标的运动特征和外观特征融入权值，可以较精确地模拟真实的跟踪场景。接着，加入虚拟结点以处理目标缺失的问题，并给出其加速版：聚合虚拟结点。最后利用最大二值整数规划求解无向图以同时获得一系列团。实验在公共数据集上进行，结果表明，该算法可以实现实时跟踪，且跟踪结果较好。     

一种改进的源节点位置隐私保护方法

当传感器网络中存在一个全局监听者的情况下,为了保护源节点的位置安全,文中提出了一种周期性发送干扰数据的算法。网络中的节点周期性地发送干扰数据包,将真实数据混淆在其中,攻击者无法判断真实数据的流向,从而保护了源节点的安全。同时,文中提出在每个节点中设立数据门限的思想,当节点中的数据超过一定门限值,将根据一定策略暂缓发送干扰数据,从而防止网络拥塞等状况的出现,同时能平衡网络的流量状况,节省资源带宽。从网络流量分析结果可知,该算法有效地保护了源节点的安全。     

Perception of delay time of head movement in using an acoustical telepresence robot: TeleHead

We built an acoustical telepresence robot, TeleHead, which has a user-like dummy head and is synchronized with the user’s head movement in real time. We are trying to clarify the effects of reproducing head movement. In this paper, we evaluate the sense of incongruity induced by the delay time in reproducing head movement. The results indicate that whether users feel the delay or not depends on their spatial perception. Therefore, acceptable delay time can be calculated from the user’s localization accuracy and head movement speed. Even under the strictest condition, i.e. high-speed head movements, using white noise makes it easier to localize sounds and roughly about 40-ms delay may be acceptable. Moreover, in conversational situations, 80-ms delay is acceptable.     

Two coding strategies for bidirectional associative memory

Enhancements of the encoding strategy of a discrete bidirectional associative memory (BAM) reported by B. Kosko (1987) are presented. There are two major concepts in this work: multiple training, which can be guaranteed to achieve recall of a single trained pair under suitable initial conditions of data, and dummy augmentation, which can be guaranteed to achieve recall of all trained pairs if attaching dummy data to the training pairs is allowable. In representative computer simulations, multiple training has been shown to lead to an improvement over the original Kosko strategy for recall of multiple pairs as well. A sufficient condition for a correlation matrix to make the energies of the training pairs be local minima is discussed. The use of multiple training and dummy augmentation concepts are illustrated, and theorems underlying the results are presented.     

Design research of the DES against power analysis attacks based on FPGA

Aiming at the DES design scheme against power analysis attacks introduced by Standart et al., an improved scheme is presented in this paper. In the improved scheme, eight dummy S-Boxes are proposed to make the power consumption of the DES S-Box logic gates constant instead of random, and it can make the same difficulties for power analysis attackers consuming 98% less memories as compared with the previous scheme. By analyzing the improved scheme in theory and using an accurate circuit simulator, the secure efficacy of the improved one is verified. The verification results indicate that the improved scheme can satisfy the practical applications against power analysis attacks, and it can be also introduced into the FPGA implementations of other cryptographic algorithms’ S-Box against power analysis attacks.     

Using genetic algorithm to optimize the dummy filling problem of the flash lamp anneal process in semiconductor manufacturing

In deep sub-micron era, many semiconductor fabrication process variations highly relate to uniformity of IC layout design. Chemical-polishing process and Flash Lamp Anneal (FLA) are two of the crucial processes aiming to increase uniformity of IC. Dummy filling is an efficient and effective Design for Manufacturability method for increasing layout uniformity by filling non-functional dummy shapes onto unoccupied area and thus reducing pattern-induced process variation. However, none are design for the thermal effects of FLA process. FLA process annealed the wafer in high temperature (1250°C) in a few milliseconds. Wafer surface emissivity determines the amount of heat absorption during FLA process. The temperature variation of FLA process induced by surface emissivity variation of IC layout results in shifts of transistors?? electrical parameters. This paper proposed to use genetic algorithm to minimize the emissivity variation of IC layout by filling a series of prescribed dummy patterns with various emissivity. The experimental results from twenty test cases show that 35% emissivity variation reductions can be achieved and moreover the observed temperature deviation during FLA is under 2.8%.     

不规则IP模块到2维NoC结构的映射方法研究

提出了一种新的基于NoC(Network on Chip)的不规则IP模块映射方法.其基本思想是把较大的IP模块分解成几个小的IP虚模型,或把几个较小的IP模块组合成一个IP虚模型,使得每个IP虚模型能映射到NoC结构的一个资源节点上.通过计算曼哈顿距离和输入/输出度,可以确定每个通信节点中缓冲区的大小.根据计算的通信代价可以对初始映射结果进行调整,从而可以避免通信拥塞,降低系统的功耗.