Spatial-temporal analysis of safety risks in trajectories of construction workers based on complex network theory

Understanding the traffic patterns of construction workers on high-risk construction sites is important for implementing behaviour-based safety management. However, safety risks in worker trajectories are a complex system with high uncertainty. This resulted in few studies focusing on analysing the spatial–temporal risk in workers' trajectories from a systematic perspective. This study designs a new framework to explore the spatial–temporal patterns of safety risks in the trajectories of construction workers based on complex network theory. First, an integrated site safety risk classification method by combining hazard sources and group trajectory distribution is developed to accurately describe the spatial distribution of site risks. Second, a new complex network chronnet is used to construct complex networks containing risk information for spatial–temporal analysis. Finally, key risk areas and risk transition patterns are identified through the analysis of network measures. The study also developed a computational program that allows the network to be constructed and analysed in real-time. The feasibility and effectiveness of the method are verified through a case study. The results show that the method can reveal the risk distribution at the micro level, and explore the risk clustering and transition features in worker trajectories at the macro level. The study allows for an accurate analysis of dynamic risk patterns in construction workers' trajectories from a systematic perspective. It can also provide theoretical and practical support for personalized and adaptive behaviour-based safety management for construction workers.     

特征挖掘与区域增强的弱监督时序动作定位

弱监督时序动作定位旨在定位视频中行为实例的起止边界及识别相应的行为。现有方法尽管取得很大进展,但依然存在动作定位不完整及短动作的漏检问题。为此,提出了特征挖掘与区域增强（FMRE）的定位方法。首先,通过基础分支计算视频片段之间的相似分数,并以此分数聚合上下文信息,得到更具有区别性的段分类分数,实现动作的完整定位;然后,添加增强分支,对基础分支定位中持续时间较短的动作提案沿时间维度进行动态上采样,进而采用多头自注意机制对动作提案间的时间结构显式建模,促进具有时间依赖关系的动作定位且防止短动作的漏检;最后,在两个分支之间构建伪标签互监督,逐步改进在训练过程中生成动作提案的质量。该算法在THUMOS14和ActivityNet1.3数据集上分别取得了70.3%和40.7%的检测性能,证明了所提算法的有效性。     

Weakly supervised temporal action localization with proxy metric modeling

Temporal localization is crucial for action video recognition. Since the manual annotations are expensive and time-consuming in videos, temporal localization with weak video-level labels is challenging but indispensable. In this paper, we propose a weakly-supervised temporal action localization approach in untrimmed videos. To settle this issue, we train the model based on the proxies of each action class. The proxies are used to measure the distances between action segments and different original action features. We use a proxy-based metric to cluster the same actions together and separate actions from backgrounds. Compared with state-of-the-art methods, our method achieved competitive results on the THUMOS14 and ActivityNet1.2 datasets.     

基于社团多特征融合嵌入表示的时序链路预测方法

时序动态网络在静态网络基础上综合了时间属性的概念,包含了网络结构的复杂性、动态性等内涵,是研究复杂网络链路预测问题的较优思维对象,因在现实世界中具有较高应用价值而备受关注。目前大部分传统方法研究对象仍局限于静态网络,存在对网络时域演化信息利用不充分、时间复杂度较高等问题。结合社会学理论,提出一种基于社团多特征融合嵌入表示的时序链路预测方法,该方法的核心思想是通过分析网络动态演化特性,在社团范围内学习节点的嵌入表示向量,融合多特征以衡量节点间连边的生成概率。利用网络集体影响力的方法对节点和连边的权值进行计算,基于集体影响的连边权值进行社团划分,将网络划分为若干个社团子图,得到基于集体影响的相似性指标。在社团范围内,利用有偏的随机游走,结合梯度优化的Skip-gram方法获取所有节点的嵌入表示向量,得到基于社团范围游走的相似性指标。融合节点的集体影响、社团范围节点的多个中心性特征和学习到的节点表示向量,得到多特征融合的相似性指标,3 种新指标都可以用于衡量节点之间形成连边的概率。对比基于移动平均、嵌入表示、图神经网络等经典时序链路预测方法,在 6 个真实数据集上的实验结果表明,所提基于社团多特征融合的方法在 AUC评价标准下取得更优的预测性能。     

基于双重注意力机制和GRU网络的短期负荷预测模型

电力负荷预测对电力系统的部署、规划和运行影响重大,但目前各输入特征对电网负荷情况影响的程度不稳定,且递归神经网络捕获负荷数据的长期记忆能力差,导致预测精度下降。提出一种基于双重注意力机制和GRU网络的预测新模型,利用特征注意力机制自主分析历史信息与输入特征间的关联关系,提取重要特征,并通过时序注意力机制自主选取GRU网络中关键时间点的历史信息,提升较长时间段预测效果的稳定性。在3个公开数据集上的实验结果表明,该模型在预测精度指标上表现良好,对比SVR、KPCA-ELM、DBN、GRU、Attention-GRU、CNN-LSTM、Attention-CNN-GRU模型预测精度分别提高了2.47、1.14、1.93、1.37、1.04、0.74、0.41个百分点。     

面向多维意象的产品设计方案评估研究

用户的感性诉求日益获得产品设计师和企业决策者的关注,为了使产品开发人员在开发新产品时能满足用户多维复合的感性诉求,提出一种结合熵处理、灰色关联分析和逼近理想解排序(technique for order preference by similarity to an ideal solution,TOPSIS)的方法,用于产品设计方案的评估。首先,在运用K-mean聚类分析和因子分析方法分别获得符合用户感性认知的典型样本和典型感性意象的基础上,运用语意差异法获得用户-意象初始评价值;其次,应用信息熵计算评估指标的权重;最后,采用灰色关联分析与TOPSIS相结合的方法实现用户多维意象特定要求下对备选设计方案的排序优选。以3D打印机设计方案优选为例,验证了方法的可行性。研究结果表明,该方法符合用户实际的复杂认知,能有效客观地对产品设计方案进行评估和优选,具有实际应用价值和指导作用。     

曲面复合材料缝合机器人自动路径规划技术

为提高复合材料预制件缝合机器人的灵活性与加工柔性,针对缝合机器人自动路径规划问题,研究了基于三维光栅扫描技术的曲面复合材料缝合机器人视觉接缝提取技术.针对三维扫描摄像机视场固定的特点,对摄像机进行标定实验,改进传统的手眼标定方法,提出一种用于求解手眼关系矩阵的四点标定新方法,建立了机器人坐标系与摄像机坐标系的位姿映射关系.经过手眼关系标定,三维光栅扫描仪采样得到的缝合预制件点云数据可转化为机器人坐标系下的点云数据坐标.进一步对采集的预制件点云数据进行接缝中心线提取,提出了一种缝合轨迹及姿态规划算法,采用三次多项式对接缝中心线进行空间曲线拟合,采用最小二乘法对缝合微切平面进行平面拟合,完成机器人操作空间前进矢量及接近矢量的计算.最后,将规划结果应用于缝合机器人的离线编程仿真与实验中.实验结果表明,该系统精度高,线迹成型良好,可以满足机器人缝合系统的要求.     

Managing sequential ambidexterity in the electronics industry: roles of temporal switching capability and contingent factors

Researchers on organisational ambidexterity have proposed several solutions to address the potential conflicts between exploration activities and exploitation activities. Unlike simultaneous ambidexterity, sequential ambidexterity – defined as temporal switching between exploration and exploitation – has not been examined fully, and the conditions under which this temporal switching can be successful are unclear. This paper proposes the concept of temporal switching capability to better understand the process by which sequential ambidexterity is executed. In addition, we hypothesise that performance effects are contingent upon firm-specific factors: a firm’s business strategy and absorptive capacity. Utilising three sources of data – a secondary database, annual reports and a survey administered to 145 firms in the electronics industry with 10-year observations – we find support for our hypotheses. The results show that the temporal switching capability positively relates to new product performance and that business strategy type and absorptive capacity have moderating effects. The results are meaningful in both theory and practice.     

Reasoning about action and change

Reasoning about change is a central issue in research on human and robot planning. We study an approach to reasoning about action and change in a dynamic logic setting and provide a solution to problems which are related to the Frame problem. Unlike most work on the frame problem the logic described in this paper is monotonic. It (implicitly) allows for the occurrence of actions of multiple agents by introducing non-stationary notions of waiting and test. The need to state a large number of frame axioms is alleviated by introducing a concept of chronological preservation to dynamic logic. As a side effect, this concept permits the encoding of temporal properties in a natural way. We compare the relative merits of our approach and non-monotonic approaches as regards different aspects of the frame problem. Technically, we show that the resulting extended systems of propositional dynamic logic preserve (weak) completeness, finite model property and decidability.