Human–machine shared control for lane departure assistance based on hybrid system theory

To improve the human–machine cooperation of lane departure assistance system (LDAS), a human–machine shared control strategy based on hybrid system theory was proposed. By considering vehicle’s discrete and continuous states and time-varying longitudinal speed, the hybrid system was formalized as hybrid automaton, and the shared control strategy was built to govern the human–machine interaction. Robust gain-scheduling energy-to-peak method was adopted to design the assistance system controller. The D-stability of the system was also studied and guaranteed by solving the linear matrix inequality (LMI). The proposed human–machine shared control method was evaluated via the co-simulation of CarSim/Simulink and the hardware-in-loop (HIL) experiment. The results showed that the proposed approach can effectively keep the vehicle in lane and a good human–machine coordination was demonstrated.     

采用信息增益率的混合入侵检测模型设计

针对现有混合入侵检测模型仅定性选取特征而导致检测精度较低的问题，同时为了充分结合误用检测模型和异常检测模型的优势，提出一种采用信息增益率的混合入侵检测模型.首先，利用信息增益率定量地选择特征子集，最大程度地保留样本信息；其次，采用余弦时变粒子群算法确定支持向量机参数构建误用检测模型，使其更好地平衡粒子在全局和局部的搜索能力，然后，选取灰狼算法确定单类支持向量机参数构建异常检测模型，以此来提高对最优参数的搜索效率和精细程度，综合提高混合入侵检测模型对攻击的检测效果；最后，通过两种数据集进行仿真实验，验证了所提混合入侵检测模型具有较好的检测性能.     

基于GF-1数据的三峡库区水体信息精细化提取

为实现江河类狭长型水体信息的精细化提取，利用GF-1卫星数据，采用支持向量机和目视解译相结合的方法对三峡库区及重庆市水体信息进行了精细化提取。使用总体分类精度、Kappa系数、错分误差、漏分误差、制图精度和用户精度等指标对库区水体信息粗提取结果进行验证分析。结果表明:4个试验区水体提取的总体分类精度均超过90%，除试验区4的Kappa系数为0.884 1以外，其余试验区均超过0.9，提取精度较高。结合目视解译的方法，在粗提取结果的基础上对各问题进行精细化处理，得到精度高、完整性好的三峡库区以及重庆市水体信息数据，可为后续该区域的精细化遥感业务开展提供有效资料。     

结合深度学习与相关反馈的遥感图像检索

目的 针对基于内容的图像检索存在低层视觉特征与用户对图像理解的高层语义不一致、图像检索的精度较低以及传统的分类方法准确度低等问题，提出一种基于卷积神经网络和相关反馈支持向量机的遥感图像检索方法。方法 通过对比度受限直方图均衡化算法对遥感图像进行预处理，限制遥感图像噪声的放大，采用自学习能力良好的卷积神经网络对遥感图像进行多层神经网络的监督学习提取丰富的图像特征，并将支持向量机作为基分类器，根据测试样本数据到分类超平面的距离进行排序得到检索结果，最后采用相关反馈策略对检索结果进行重新调整。结果 在UC Merced Land-Use遥感图像数据集上进行图像检索实验，在mAP（mean average precision）精度指标上，当检索返回图像数为100时，本文方法比LSH（locality sensitive Hashing）方法提高了29.4%，比DSH（density sensitive Hashing）方法提高了37.2%，比EMR（efficient manifold ranking）方法提高了68.8%，比未添加反馈和训练集筛选的SVM（support vector machine）方法提高了3.5%，对于平均检索速度，本文方法比对比方法中mAP精度最高的方法提高了4倍，针对复杂的遥感图像数据，本文方法的检索效果较其他方法表现出色。结论 本文提出了一种以距离评价标准为核心的反馈策略，以提高检索精度，并采用多距离结合的Top-k排序方法合理筛选训练集，以提高检索速度，本文方法可以广泛应用于人脸识别和目标跟踪等领域，对提升检索性能具有重要意义。     

异型中间包结构优化对帘线钢夹杂物的影响

为了解决帘线钢中因大颗粒夹杂物导致的拉拔断裂问题，通过水模型研究结果对异型中间包结构改进优化，并将优化异型中间包结构进行工业试验。结果表明，挡墙优化结构后,去除夹杂物的效果明显提高，采用优化的挡墙中间包生产的帘线钢铸坯中夹杂物尺寸不大于10 μm比例为87.95%，平均尺寸为6.7 μm；大颗粒的高MgO类和高SiO2类夹杂物已得到有效去除，盘条中B细不大于0.5级比例达到97.56%。     

[成形过程的数据挖掘与深度学习方法]基于机器学习的管材弯曲回弹有效预测与补偿

采用基于优化的误差反向传播(BP)神经网络的机器学习算法建模，提出了考虑材料参数、几何参数等多因素的弯管回弹精确预测和高效控制方法。该方法通过引入非线性惯性权重及遗传算法的杂交算子，改进了粒子群优化(PSO)算法，进而通过改进的PSO算法对BP神经网络进行优化，构建了基于改进的PSO-BP神经网络机器学习回弹预测和补偿模型。以多种规格的铝合金数控弯管构件为对象，将实际生产中不同规格、批次、成形参数下回弹数据作为训练样本，实现了所建机器学习预测模型的应用验证。所建模型获得的预测结果平均相对误差为6.3%，与未优化的BP神经网络等传统模型相比，预测精度最大提高了18.5%，计算时间可从1.5 h缩短至300 s，同时实现了回弹预测与补偿精度以及计算效率的显著提高。     

基于机器视觉的玻璃缺陷检测

随着科学技术的进步,高端显示屏产品对平板玻璃的质量要求越来越高,玻璃的表面缺陷检测技术也因此备受关注。传统的人眼检测方法工作量大且准确率低,已经无法满足生产实际要求。研究了一种基于机器视觉的玻璃质量检测系统,采用先进的CCD成像技术和背光式照明获取图像,用MATLAB图像处理工具箱对采集到的图像进行灰度值化、滤波降噪和阈值分割处理,实现对缺陷区域的特征提取和识别。最后用BP神经网络对玻璃表面的三种缺陷进行分类,该神经网络识别的平均误差率为9.84%,表明此检测方法具有一定的应用价值。     

基于机器视觉的航空铆钉尺寸检测技术

航空铆钉的尺寸偏差会增加航空事故发生的可能性。为了实现航空铆钉尺寸检测的高精度与智能化,提出一种基于机器视觉的航空铆钉尺寸检测方法。通过边缘检测算法对铆钉进行自动识别与检测,可减少航空铆钉尺寸的误差。基于航空铆钉产品规范,分析得出各项尺寸参数要求与事件流程,使用OpenCV作为开发平台,采用阈值分割算法与分水岭算法进行图像分割;同时,通过平滑处理、形态学、轮廓提取算法完成图像处理。实测结果表明,在±0.005mm精度下,铆钉尺寸检测结果与实际值的吻合度达到90%以上。     

Modeling of quantitative relationships between physicochemical properties of active pharmaceutical ingredients and tensile strength of tablets using a boosted tree

Objectives: The aim of this study was to explore the potential of boosted tree (BT) to develop a correlation model between active pharmaceutical ingredient (API) characteristics and a tensile strength (TS) of tablets as critical quality attributes.

Methods: First, we evaluated 81 kinds of API characteristics, such as particle size distribution, bulk density, tapped density, Hausner ratio, moisture content, elastic recovery, molecular weight, and partition coefficient. Next, we prepared tablets containing 50% API, 49% microcrystalline cellulose, and 1% magnesium stearate using direct compression at 6, 8, and 10?kN, and measured TS. Then, we applied BT to our dataset to develop a correlation model. Finally, the constructed BT model was validated using k-fold cross-validation.

Results: Results showed that the BT model achieved high-performance statistics, whereas multiple regression analysis resulted in poor estimations. Sensitivity analysis of the BT model revealed that diameter of powder particles at the 10th percentile of the cumulative percentage size distribution was the most crucial factor for TS. In addition, the influences of moisture content, partition coefficients, and modal diameter were appreciably meaningful factors.

Conclusions: This study demonstrates that BT model could provide comprehensive understanding of the latent structure underlying APIs and TS of tablets.     

Modeling and surface texturing on surface roughness in machining LaPO4–Y2O3 composite

In this work, lanthanum phosphate with a 20% yttria (LaPO₄/Y₂O₃) composite prepared by an Aqueous Sol–Gel process is machined using an Abrasive Water Jet Machine (AWJM). The machinability of this composite is studied by varying the input parameters namely Jet Pressure (JP), Stand-Off Distance (SOD), and Traverse Speed (TS) on Surface finish. Garnet of 80 mesh size is used as an abrasive with a flow rate of 85?g/min. The microstructural characterization study reveals the presence of new element YPO₄. This element enhances the machinability and reduced porosity in the composite. Microscopy examinations on the machined surface reveals that partial overlapping at low JP, poor surface finish at high JP and SOD, forged deficiency at maximum SOD and TS. The minimum levels of all input parameters are influenced to obtain acceptable Ra. Atomic Force Microscopy (AFM) on the kerf surface shows micro wear track and peaks. The Multiple Regression Analysis (MRA) is developed for Ra to check the adequacy. From the Analysis of Variance (ANOVA), SOD has a significant effect on Ra with a contribution of 53%. The influence of JP and TS on Ra is found to be 31% and 15%, respectively.