Effective pedestrian detection using deformable part model based on human model

Recently, pedestrian detection systems have become an important technology in the development of the advanced driver assistance system (ADAS) for the autonomous car. The histogram of oriented gradients (HOG) is currently the most basic algorithm for detecting pedestrians, but it treats the entire body of the pedestrian as one single feature. In other words, if the entire body of the pedestrian is not visible, the detection rate under HOG decreases markedly. To solve this problem, we propose a detection system using a deformable part model (DPM) that divides the pedestrian data into two parts using a latent support vector machine (SVM)-based machine-learning technique. Experimental results show that our approach achieves better performance in a detection system than the existing method. In practice, there are many occlusions in the environment in front of the vehicle. For example, the surrounding transport facilities, such as a car or another obstacle, can occlude a pedestrian. These occlusions can increase the false detection rate and cause difficulties during the detection process. Our proposed method uses a different approach and can easily be applied in real-world scenarios, regardless of occlusions.

     

Calibrating a social force based model for simulating personal mobility vehicles and pedestrian mixed traffic

Personal mobility vehicles (PMVs), such as the Segway, have recently gained remarkable popularity as an alternative transport mode for short-distance trips in indoor and outdoor settings. Before allowing them on shared sidewalks, where the pedestrian and cyclist demand is high, interactions between PMV riders and other shared space users should be properly understood. Further, the designs of shared sidewalks and implementation policies should also be evaluated. Calibrated microscopic simulation tools could facilitate such purposes. This study aims to explore the applicability of a social force based microscopic simulation model, which was originally used to simulate pedestrian movements and interactions, for Segway and pedestrian mixed traffic. The parameters of the model are calibrated with data collected through controlled experiments under different Segway–pedestrian interaction scenarios. Lateral and longitudinal avoidance distances measured from trajectory data collected in a different controlled experiment was used to validate the model for a Segway rider avoiding a pedestrian. The findings of this study suggest that, with proper calibration, the social force model can potentially be used to simulate Segway-like PMVs and pedestrian mixed traffic.     

动态服务的数据交换模型研究*

数据交换是解决信息孤岛问题的实用方法,但如何实施灵活有效的数据交换一直是研究的难点.基于此,提出了动态服务的数据交换方式来克服原有单一、固定的交换缺点.依据交换系统的现状,首先给出了静态服务库、动态服务库和静态交换系统、动态交换系统的概念;然后提出了动态服务的交换模型,并给出了交换的方法、原则和实现过程.最后,结合具体的应用实例,验证了模型的灵活性和有效性.     

A genetic fuzzy system to model pedestrian walking path in a built environment

A study on the pedestrian’s steering behaviour through a built environment in normal circumstances is presented in this paper. The study focuses on the relationship between the environment and the pedestrian’s walking trajectory. Owing to the ambiguity and vagueness of the relationship between the pedestrians and the surrounding environment, a genetic fuzzy system is proposed for modelling and simulation of the pedestrian’s walking trajectory confronting the environmental stimuli. We apply the genetic algorithm to search for the optimum membership function parameters of the fuzzy model. The proposed system receives the pedestrian’s perceived stimuli from the environment as the inputs, and provides the angular change of direction in each step as the output. The environmental stimuli are quantified using the Helbing social force model. Attractive and repulsive forces within the environment represent various environmental stimuli that influence the pedestrian’s walking trajectory at each point of the space. To evaluate the effectiveness of the proposed model, three experiments are conducted. The first experimental results are validated against real walking trajectories of participants within a corridor. The second and third experimental results are validated against simulated walking trajectories collected from the AnyLogic® software. Analysis and statistical measurement of the results indicate that the genetic fuzzy system with optimised membership functions produces more accurate and stable prediction of heterogeneous pedestrians’ walking trajectories than those from the original fuzzy model.     

基于原子力显微镜动力学模型的新型接触式扫描成像策略

面向高速高精度原子力显微镜的扫描成像策略是当今微纳米控制领域的热点问题。本文设计了一种新型的原子力显微镜接触式扫描算法,该算法充分挖掘了AFM压电陶瓷运动平台和微悬臂的动力学特性,通过对微悬臂偏转角进行跟踪控制以达到AFM高速高精度成像的目的。在动力学建模方面,本文提出了一种新颖的基于AFM压电陶瓷运动平台动态方程及微悬臂动态特性的复合动力学模型,该模型所构建的动力学系统可以通过压电陶瓷的输入电压对微悬臂偏转角进行精确控制。在此基础上,本文设计了一种新型的图像融合算法,利用AFM正反向扫描得到的两幅图像进行融合匹配,能够较好地消除AFM控制周期暂态特性所带来的图像畸变,提高AFM扫描成像准确度。本文所提出扫描成像策略的有效性在AFM实验平台上进行了验证,实验结果表明此成像策略能够在较高扫描频率下得到较为理想的成像效果。最后通过对比实验将这种方法与传统接触式扫描算法的成像效果进行了比较,结果表明对比传统方法本文所提出的AFM成像策略可以更好地保证成像速度和精确度。     

Outlier detection based on a dynamic ensemble model: Applied to process monitoring

This paper focuses on outlier detection and its application to process monitoring. The main contribution is that we propose a dynamic ensemble detection model, of which one-class classifiers are used as base learners. Developing a dynamic ensemble model for one-class classification is challenging due to the absence of labeled training samples. To this end, we propose a procedure that can generate pseudo outliers, prior to which we transform outputs of all base classifiers to the form of probability. Then we use a probabilistic model to evaluate competence of all base classifiers. Friedman test along with Nemenyi test are used together to construct a switching mechanism. This is used for determining whether one classifier should be nominated to make the decision or a fusion method should be applied instead. Extensive experiments are carried out on 20 data sets and an industrial application to verify the effectiveness of the proposed method.     

基于视觉注意力模型的红外人体图像兴趣区域提取*

由于红外人体图像所处背景的复杂性,使用阈值分割定位红外人体图像往往把大量的非人体图像一起分离出来,这给后续工作带来了巨大的计算量.研究了红外人体图像的典型特点和人眼视觉判断人体轮廓图像的敏感特征,使用该特征建立了简单的人眼视觉对人体的注意力模型,并构建特征判别概率函数,对待定目标进行概率的估算,最终使用该模型快速地剔除非人体红外图像的干扰,并准确地定位人体兴趣区域.     

基于离心原理的力传感器动态标定系统辨识与分析

针对静态校准过的传感器用来测量动态力时,可能会出现较大误差的情况,设计了基于水平正弦力加载的力传感器动态标定装置,动态力由旋转的偏心轮产生。通过分析建立了系统的动力学模型,使用最小二乘辨识方法得到了系统传递函数模型参数,并在多组频率下分别验证模型的准确性。实验结果表明:建立并辨识得到的系统模型可以较好地逼近实际物理系统,并且根据频率改变参数得到的模型比固定参数模型具有更高的精度。     

改进R-FCN模型的小尺度行人检测

目的 为了有效解决传统行人检测算法在分辨率低、行人尺寸较小等情境下检测精度低的问题,将基于区域全卷积网络（region-based fully convolutional networks,R-FCN）的目标检测算法引入到行人检测中,提出一种改进R-FCN模型的小尺度行人检测算法。方法 为了使特征提取更加准确,在ResNet-101的conv5阶段中嵌入可变形卷积层,扩大特征图的感受野;为提高小尺寸行人检测精度,在ResNet-101中增加另一条检测路径,对不同尺寸大小的特征图进行感兴趣区域池化;为解决小尺寸行人检测中的误检问题,利用自举策略的非极大值抑制算法代替传统的非极大值抑制算法。结果 在基准数据集Caltech上进行评估,实验表明,改进的R-FCN算法与具有代表性的单阶段检测器（single shot multiBox detector,SSD）算法和两阶段检测器中的Faster R-CNN（region convolutional neural network）算法相比,检测精度分别提高了3.29%和2.78%;在相同ResNet-101基础网络下,检测精度比原始R-FCN算法提高了12.10%。结论 本文提出的改进R-FCN模型,使小尺寸行人检测精度更加准确。相比原始模型,改进的R-FCN模型对行人检测的精确率和召回率有更好的平衡能力,在保证精确率的同时,具有更大的召回率。     

基于行为监控和数据挖掘的动态信任模型*

实体之间的信任关系发生在一定的上下文中,其信任值与影响信任值的多个行为属性之间的关系复杂而且多变,很难用一个一成不变的函数去描述。根据软件传感器监测到的历史行为数据和目标信任值,利用logistic回归分析方法和成对分类法对行为属性与信任值之间的关系模式进行自适应的数据挖掘与知识发现,不需要任何的先验知识和主观假设,从而有效解决了信任值计算的动态性和客观性等问题。实验结果表明,与已有模型相比,该模型的平均误差率和计算效率都有所提高,能够快速有效地分类出实体间的信任等级。     

改进社会力的双向行人流模型研究

在行人运动和行人行为观察的基础上,结合社会力模型,把行人运动看作一种自驱动的个体在连续空间中移动的过程。为了解决行人间相互作用力的不足,通过引入行人运动感知域,提出一种行人碰撞避免的解决机制,这种机制能较好解决行人间的相互碰撞行为,防止行人间出现大面积的堵塞现象。仿真实验结果显示改进后的模型能再现正常条件下行人流基本图,并很好地再现了行人流“通道形成”自组织现象。     

基于PQCR-PSL的桥梁振动无线监测系统

基于压电式石英晶体谐振器与平面螺旋电感串联结构(PQCR-PSL),结合现场可编程门阵列(FPGA)、无线传感器网络(WSNs)和C#等技术,将一种结构简单、以频率为输出的数字化石英晶体电感传感器应用到桥梁振动等低频测量中,并以此为传感器探头搭建了整套无线监测系统.系统主要由传感器探头、FPGA系统、ZigBee网络、C#位机4部分组成,具有结构简单、稳定性好、数字化输出等特点,实现了对低频振动信号的非接触式无线实时监测.     

基于同步监测的锚杆受力监测系统设计

针对现有锚杆受力监测方法存在的结构复杂、测点单一、安装不便等问题,设计了基于同步监测的锚杆受力监测系统。该系统选用扩散硅压力传感器采集巷道某一断面的锚杆轴向力,可同时获取8根锚杆的受力数据;将采集的数据进行分压滤波、AD转换后,存储并传输。实验结果表明,该系统实现了对巷道某一断面多根锚杆、锚索的高采样率同步监测,测量精度可以满足断面锚杆受力同步监测的要求。     

Adaptive dynamic walking of a quadruped robot using a neural system model

We are trying to induce a quadruped robot to walk dynamically on irregular terrain by using a neural system model. In this paper, we integrate several reflexes, such as a stretch reflex, a vestibulospinal reflex and extensor/flexor reflexes, into a central pattern generator (CPG). We try to realize adaptive walking up and down a slope of 12°, walking over an obstacle 3 cm in height, and walking on terrain undulation consisting of bumps 3 cm in height with fixed parameters of CPGs and reflexes. The success in walking on such irregular terrain in spite of stumbling and landing on obstacles shows that the control method using a neural system model proposed in this study has the ability for autonomous adaptation to unknown irregular terrain. In order to clarify the role of a CPG, we investigate the relation between parameters of a CPG and the mechanical system by simulations and experiments. CPGs can generate stable walking suitable for the mechanical system by receiving inhibitory input as sensory feedback and generate adaptive walking on irregular terrain by receiving excitatory input as sensory feedback. MPEG footage of these experiments can be seen at: http://www.kimura.is.uec.ac.jp.     

桌面网格环境下遥感数据动态存储模式研究

基于网格体系结构节点资源能够被共享并被协同使用的概念,设计在分布式节点上实现数据存储和传递的网格数据存储系统.该系统允许网格用户在本地将数据上传到网格,网格管理节点负责对参与共享存储资源的节点进行管理,并将预定大小的数据分配到相应节点存储,同时响应网格用户的请求,使用基于Hash表的路由信息,找到对应请求网格数据的最佳路径,并激活网格线程,实现网格数据在节点间的完整传递.基于Alchemi网格中间件和.NET框架对遥感数据而进行的开发和应用表明,桌面网格动态存储是一个可行的网格计算应用.     

基于多传感器数据融合的室内行人航位推算法

针对惯性测量单元(IMU)本身存在测量漂移,很难获得精确的室内行人轨迹的问题,提出了使用多个传感器信息融合的方案,包括IMU和视频摄像头,并结合卡尔曼滤波和零速检测算法进行参数优化,以提高行人运动轨迹的精度.仿真结果表明:算法可以有效降低行人轨迹的误差.     

改进星型级联可形变部件模型的行人检测

目的 行人检测是计算机视觉和模式识别领域的研究热点与难点,由于经典的可形变部件模型(DPM)检测速度太慢,引入PCA降维的星型级联检测可形变部件模型(casDPM)相比较于DPM模型检测速度虽然有了很大提升,但在应用于行人检测时,出现检测精度较低、平均对数漏检率较高的情况,为了更加准确地对行人进行检测,提出了一种改进casDPM模型的行人检测方法。方法 首先利用对象度量方法获取目标候选区域,结合目标得分信息得到casDPM模型低分检测区域的置信度,在设定的阈值上保留检测窗口;然后针对casDPM模型原有非极大值抑制(Nms)算法只利用单一的面积信息,造成误检数较高的情况,提出了利用检测窗口的得分信息进行改进;最后将两种方法结合起来,提出了融合的cas-WNms-BING模型。结果 采用本文方法在INRIA数据集上进行检测,实验结果表明该方法对于行人形变、背景特征复杂及遮挡现象具有较强的鲁棒性,相比casDPM模型,本文提出的方法平均精度(AP)可以提高1.74%,平均对数漏检率可以降低4.45%。结论 提出一种改进星型级联可形变部件模型,取得一定的研究成果,在复杂的背景下,能够有效地进行行人检测,主观视觉感受和客观实验评价指标都表明该方法可以有效提升模型行人检测效果。但是,星型级联可形变部件模型训练及检测效率仍有待提高,需进一步对模型存在的一些局限性进行深入研究。     

Constructing dynamic model for thermal flow forecasting based on experimental data

We consider an identification algorithm for a dynamic object. The algorithm allows for choosing any rows from the block of the transformed empirical input-output data. Moreover, the algorithm enables passing to a parameter space and involving a priori information on the domain of parameter existence to improve accuracy. These features serve to construct a model of the dynamic object with a smaller forecasting error. The algorithm is applied for mathematical modeling of a fuel assembly in a nuclear reactor based on experimental heat transfer data.     

基于团块模型的行人阴影抑制算法

为有效解决行人阴影对行人跟踪产生的影响,提出一种基于团块模型的阴影抑制算法。该算法利用团块结构作为行人模型,通过计算团块的矩特征和团块垂直直方图,借助几何的方法获得疑似阴影区域。然后,将该疑似阴影区域的灰度、方向、中心位置作为参数对整个前景区域进行阴影建模,使得阴影区域更精确。最后,从行人模型中剔除阴影团块达到阴影抑制的效果。经过视频序列测试,本文算法对不同运动方向、不同数量的行人均有良好的阴影抑制效果,表明了该设计方案的有效性和准确性。     

Building a three-dimensional dynamic model of left cardiac ventricle from ultrasonic data

A three-dimensional dynamic reconstruction algorithm of the left ventricle of patient??s heart from ultrasonic echocardiography data is developed. The presented algorithm takes into account the longitudinal and transverse velocities of ventricle walls to determine the motion vectors of the ventricular walls. Two longitudinal ultrasonic sections (four-chamber position and two-chamber position) and two transverse ultrasonic sections (at a level of basal segments and at a level of papillary muscles) are used for the reconstruction.