基于峰宽修正的玻璃药瓶内氧气残留浓度的检测研究

玻璃药瓶中的氧气残留对瓶中药品的无菌特性造成了严重的威胁。采用波长调制光谱（Wavelength Modulated Spectrum, WMS）技术解调得到的二次谐波信号峰高值作为氧浓度反演的基础。然而,在用二次谐波信号测量气体浓度时,由于调制深度的变化会导致二次谐波峰值的变化,这通常会给系统带来误差,进而降低浓度的反演精度。而调制深度受调制电流波动、温度和压强变化等因素的影响不能直接计算获得进而修正。针对这一问题,本文成功地将调制深度与二次谐波峰高的关系转换为二次谐波峰宽和峰高的关系。然后,利用峰宽校正后的谐波峰高对气体浓度进行反演。初步实验表明,利用峰宽校正后的谐波峰高对瓶内气体浓度进行预测的准确性提高了2.1%,且系统的整体鲁棒性也得以提升。本文提出的校正方法不需要系统参数信息与气体成分信息,可以直接从谐波信号本身出发对调制深度进行校正,十分适合于工业现场的在线氧气浓度预测。     

基于改进的SSD监理目标检测研究

针对装饰装修工程中由人工验收带来的诸多问题,文中提出了一种改进的SSD算法并将其应用于监理工作来代替人工验收,推动智能监理的实现。由于SSD算法存在对同一目标复检以及小目标检测效果欠佳等问题,故文中利用DPN网络替换基础特征提取网络VGG16。DPN结合了Resnet和Densenet的优点,具有更好的特征提取能力。通过加权FPN融合特征图,突出不同层特征图的贡献,丰富用于预测的特征图语义。利用深度可分离卷积降低模型的参数量,提高算法的推理速度。实验对比发现,改进后模型的平均精度提升了3.47%,对小数目检测平均精度的提升可达15%,证明新模型在监理目标检测任务中效果良好。     

Travel time models for deep-lane unit-load autonomous vehicle storage and retrieval system (AVS/RS)

Autonomous vehicle storage and retrieval systems use vehicles that move horizontally along rails within the storage racks, while vertical movements are provided by lifts. The solution proposed in this paper addresses a particular system configuration that works with multiple deep storage lanes that are widely used in the food and beverage industry, characterised by large volumes of products of limited variety. The generic deep lane is single item, i.e. one stock keeping unit, and single batch, i.e. one production lot, thereby affecting the performance of the system in terms of storage capacity utilisation and throughput. Determining the number and depth of the lanes is crucial to aid the design and control of such a storage system. The aim of this paper was to support the design of AVS/RSs though a set of original analytic models for the determination of the travelled distance and time for single-command and dual-command cycles given alternative layout configurations. The models are validated by simulation and exemplified with a real-warehousing case study. The paper presents useful guidelines for the configuration of the system layout including the determination of the optimal shape ratio and the length of the lanes.     

基于Kinect传感器的移动机器人环境检测及行为学习

研究了一种基于深度图像和强化学习算法的移动机器人导航行为学习方法。该方法利用机器人装配的Kinect传感器检测工作环境信息,然后对获取的深度图像数据和视频图像进行处理、融合和识别,并由此构建机器人任务学习的状态空间,最终利用强化学习方法实现移动机器人的导航任务的自学习。该方法的有效性通过实验得到验证。实验表明,该方法能够使机器入具有较强的环境感知能力,并能够通过自学习的方式掌握行为能力。     

Evaluation of rutting potential for asphalt concrete mixes containing copper slag

Copper slag (CS) is a by-product of the copper extraction process, which can be used as coarse and/or fine aggregate in hot mix asphalt (HMA) pavements. This study used CS as a replacement of the fine aggregate with a percentage of up to 40% by total aggregate weight. The objective of this study was to evaluate the effect of CS on the rutting potential of the asphalt concrete mix using two methods. One method is based on the Dynamic modulus |E*| testing result. Actual pavement temperature data from a test section were used with the developed |E*| master curves. EverStressFE finite element program was used to perform a linear elastic load-deformation analysis for a pavement section and to determine the vertical resilient strain in a 40-mm HMA surface layer. The M-E PDG permanent deformation model was used with and Excel Visual Basic for Applications code to predict the accumulated rutting for different CS mixes for 10 million ESALs. The other method used the data from the flow number (FN) test. Based on the |E*| approach, the results indicated that adding 5% CS in the mix increased the predicted rutting from 0.59 to 0.98 mm at 10 million ESALs (increase by 68%). When 40% CS was used, rutting increased by more than 700% compared with the control mix. After analysing the FN results with the Francken model, the results indicated a decrease in FN as CS content is increased, indicating higher rutting potential. The decrease in FN ranged from 9% for 5% CS to 95% for 40% CS. The mixes containing up to 10% CS satisfied the minimum FN criteria for rutting. A calibration process for the M-E PDG distress prediction models that allows the use of waste and by-product materials such as CS should be considered in the future.     

Optimization of interior roller burnishing process for improving surface quality

Improving the surface characteristics of roller burnishing processes is one of effective approaches to decrease the machining costs and time. This paper systematically investigates the nonlinear relationships between machining parameters and surface characteristics, including surface roughness (R_a), surface hardness (H), and hardness depth (HD) of the interior roller burnishing using response surface method (RSM) model. Three process parameters considered include spindle speed S, feed rate F, and burnishing depth D. A set of physical experiments was carried out with AISI 1045 steel on a computer numerical control (CNC) milling machine using the roller burnishing tool. The target of the current complex optimization is to enhance the surface hardness and hardness depth, while the surface roughness is considered as the constraint. Finally, an evolutionary algorithm entitled archive-based micro genetic algorithm (AMGA) was used to generate a set of feasible optimal solutions and determine the best machining conditions. The results show that an appropriate trade-off solution can be drawn with regard to the low surface roughness and high the surface hardness as well as hardness depth. Furthermore, the integration of RSM model and AMGA can be considered as a powerful approach for modeling and optimizing interior roller burnishing processes.     

Rational Filters for Passive Depth from Defocus

A fundamental problem in depth from defocus is the measurement of relative defocus between images. The performance of previously proposed focus operators are inevitably sensitive to the frequency spectra of local scene textures. As a result, focus operators such as the Laplacian of Gaussian result in poor depth estimates. An alternative is to use large filter banks that densely sample the frequency space. Though this approach can result in better depth accuracy, it sacrifices the computational efficiency that depth from defocus offers over stereo and structure from motion. We propose a class of broadband operators that, when used together, provide invariance to scene texture and produce accurate and dense depth maps. Since the operators are broadband, a small number of them are sufficient for depth estimation of scenes with complex textural properties. In addition, a depth confidence measure is derived that can be computed from the outputs of the operators. This confidence measure permits further refinement of computed depth maps. Experiments are conducted on both synthetic and real scenes to evaluate the performance of the proposed operators. The depth detection gain error is less than 1%, irrespective of texture frequency. Depth accuracy is found to be 0.51.2% of the distance of the object from the imaging optics.     

物联网安全接入网关的设计与实现

文章通过分析物联网发展过程中存在的安全问题,提出了如何通过物联网安全接入网关,构建接入安全、管理可控的物联网系统架构,最后详细设计了物联网安全接入网关的功能组成和安全技术要求。     

A novel three-tiered visualization approach for firewall rule validation

Firewall is one of the most critical elements of the current Internet, which can protect the entire network against attacks and threats. While configuring the firewalls, rule configuration has to conform to, or say be consistent with, the demands of the network security policies such that the network security would not be flawed. For the security consistency, firewall rule editing, ordering, and distribution must be done very carefully on each of the cooperative firewalls, especially in a large-scale and multifirewall-equipped network. Nevertheless, a network operator is prone to incorrectly configure the firewalls because there are typically thousands or hundreds of filtering/admission rules (i.e., rules in the Access Control List file, or ACL for short), which could be set up in a firewall; not mentioning these rules among firewalls affect mutually and can make the matter worse. Under this situation, the network operator would hardly know his/her misconfiguration until the network functions beyond the expectation. For this reason, our work is to build a visualized validation system for facilitating the check of security consistency between the rule configuration of firewalls and the demands of network security policies. To do so, the developed validation system utilizes a three-tiered visualization hierarchy along with different compound viewpoints to provide users with a complete picture of firewalls and relationships among them for error debugging and anomaly removal. In addition, in this paper, we also enumerate the source of security inconsistency while setting ACLs and make use of it as a basis of the design of our visualization model. Currently, part of the firewall configuration of our campus network has been used as our system's input to demonstrate our system's implementation.     

融合深度信息的视觉注意计算模型

针对Itti模型在特征选择上的局限性,借鉴心理学中有关视觉注意的研究成果,提出一种融合深度信息的视觉注意计算模型。该模型在基于图像分割的自适应立体匹配基础上提取深度特征,与亮度、方向、颜色特征相结合,实现空间显著性度量,并采用侧抑机制和WTA机制得到注意焦点。实验结果表明,新模型能较好地反映空间立体视觉信息对注意的影响,使计算结果能更符合人类视觉。