基于方向可变H aar特征和双曲线模型的车道线检测方法

针对快速路车道线易受多种因素影响而较难检测的问题,提出了一种基于方向可变 Haar特征和双曲线模型的分布式车道线检测方法。首先对车载摄像头进行标定,确定图像中车道平面消失线的位置,将车道平面消失线以下部分的下2/3区域作为感兴趣区域Ⅰ。利用边缘分布函数获得感兴趣区域Ⅰ内车道线直线模型倾角,再采用方向可变 Haar 特征提取边缘特征点并拟合车道线直线模型,利用直线模型参数进一步确定感兴趣区域Ⅱ。提出一种单方向搜索算法,提取边缘特征点并利用双曲线模型拟合获取完整的车道线模型。通过约10000幅实际道路图片对车道线检测方法进行验证。验证结果表明：检测方法能很好地实现多种环境下的车道线检测,在晴好天气检测率为99．9％,不良天气检测率为99．7％。     

二次半定规划的原始对偶预估校正内点算法

将半定规划（Semidefinite Programming,SDP）的内点算法推广到二次半定规划（QuadraticSemidefinite Programming,QSDP）,重点讨论了AHO搜索方向的产生方法.首先利用Wolfe对偶理论推导得到了求解二次半定规划的非线性方程组,利用牛顿法求解该方程组,得到了求解QSDP的内点算法的AHO搜索方向,证明了该搜索方向的存在唯一性,最后给出了求解二次半定规划的预估校正内点算法的具体步骤,并对基于不同搜索方向的内点算法进行了数值实验,结果表明基于NT方向的内点算法最为稳健.     

����̰�������㷨��ͣ��λָ�������о�

停机位指派是机场运营管理的关键活动之一,合理高效的停机位指派,除有利于提高机场运行效率和旅客满意率水平之外,对降低航空公司地面运行成本也有重要意义. 本文从航空公司、机场及旅客三者综合效能最大化角度出发,以飞机地面滑行距离、机场停机位空闲时间及旅客步行距离总成本最小为目标函数,构建机场停机位指派模型,设计了指派模型的贪婪禁忌搜索算法. 该算法首先采用贪婪算法产生一个初始指派方案,然后采用禁忌搜索算法在初始指派方案的基础上进行逐步优化. 算例研究验证了模型的正确性和算法的有效性,该模型和算法可满足国内大型机场停机位指派实际运作需要.     

Train Control to Reduce Delays upon Service Disturbances at Railway Junctions

Train delay is one of the most important indexes to evaluate the service quality of the railway. Because of interactions among different trains, a delayed train may hinder subsequent trains scheduled on other lines at junction area. Trains may be forced to stop or decelerate, which consequently leads to the loss of run-time. This paper proposes a time-saving control method to recover the train delay as much as possible. In the proposed method, golden section search is adopted to identify the optimal train speed at the time of signal upgrade in the conflicting area, which enables the train to depart from the conflicting area as soon as possible. A heuristic method is then developed to attain the train control scheme to realize the optimal train speed at the time of signal upgrade in the conflicting area. Simulation study indicates that the proposed method enables the train to across the conflicting area with smaller run-time and therefore leads to larger delay reduction, in comparison with the traditional maximum speed control method and the green wave control method.     

船舶电力系统中基于BFS的开关动作排序控制方法

为了解决现有船舶电网保护方法难以达到复杂网络对保护选择性要求的弱势,提出基于广度优先搜索法(Breadth First Search)的开关动作排序法。该方法通过控制开关的延时长短,让上下级开关的时间设定值相互配合,达到保护选择性要求。最后用3个实例对比原有方法证明了新方法的优势和有效性。     

驾驶员视觉搜索模式模糊聚类评价方法

针对用单一眼动参数评价驾驶员视觉搜索模式存在片面性和较大误差等不足,提出运用模糊理论,将驾驶员的注视持续时间、水平方向视角、垂直方向视角、扫视幅度和扫视速度5个主要表征参数综合起来进行评价的方法。通过真实交通环境中的实车试验,记录驾驶员行驶过程的眼动数据,建立模糊相似矩阵和模糊等价矩阵,依据聚类样本驾驶员的驾驶里程确定最佳阈值,将样本驾驶员视觉搜索模式分为4类,并以每类驾驶员视觉搜索模式特征为判别标准,运用择近原则对其他驾驶员的视觉搜索模式进行识别和评价。结果表明:所提出的驾驶员视觉搜索模式模糊聚类评价方法有效且可靠。     

某型航空反潜鱼雷搜索概率问题仿真研究

几种典型航空反潜鱼雷的搜索路径基本上是圆周形,其搜索区域可以看作是一个小圆.对鱼雷的搜索过程进行合理的假设简化,建立其搜索模型,并对发现潜艇这个随机事件作统计试验.分析结果可知,提高鱼雷搜索半径,降低鱼雷搜索前的延迟时间,可以显著改善鱼雷的发现概率;反潜机对目标的定位误差和鱼雷的投置偏差,都会影响到搜索概率;从不同角度攻击潜艇,潜艇采取的规避方式,也会影响到捕捉概率,这些结果是值得作战指挥员关注的.     

提升新能源汽车专业服务地方产业发展能力的探索与实践

专业服务地方产业发展的能力是职业院校建校水平和人才培养质量的反映,随着河南省首张纯电动"牌照"落户三门峡,三门峡市新能源汽车产业对相关高技能专业人才培养提出了强劲需求,但是,目前职业院校还存在主动对接汽车支柱产业不足,汽车专业设置与产业需求脱节等问题。产教融合、校企合作的办学模式是职业教育改革的成功经验,地方职业院校应立足区域产业的转型发展,建立"岗位主导·项目带动"的人才培养模式,以汽车专业群建设为引领,从专业结构优化调整、人才规格紧贴行业、推进校企对接、加强专业课程改革力度、打造"双师型"教师队伍、推行校企"双主体"育人等方面入手,培养工业企业急需的高技能人才,助力工业转型发展。     

2750TEU集装箱船局部振动评估

针对2 750 TEU集装箱船,在船体总振动评估的基础上,采用经验公式及有限元法,对液舱及上层建筑的局部振动进行详细评估,并根据不同区域的计算结果采取相应的改进措施.     

Determination of t8/5 cooling times for underwater local dry welding of steel

Knowledge of thermal history is the basic condition for studying the structure - properties of welded joints. The determinant of thermal history is the thermal cycle, whose in-situ measurements are still a big challenge. Water as the welding environment complicates this issue even more. The article presents a method to determine an equation for calculating t_8/5 cooling times for underwater gas metal arc welding of unalloyed steels using the local dry cavity method. The work uses the contact method of temperature measurements with the use of thermocouples to obtain the temperature changes of the points of welded joints covered by the thermal field. On this basis, the values of the t_8/5 cooling times were determined. A regression analysis of the Response Surface Method was used to determine the equation, which resulted in a second-order model with interactions. Statistically significant factors were determined (thickness of welded elements and heat input value) and the model veracity was confirmed as the assumptions of normality and homogeneity of variance (homoscedasticity) of the residuals were met. The t_8/5 cooling time values depend on the thickness of the welded elements and heat input value in a nonlinear way. In terms of the test conditions, the cooling times ranged from 3.0 to 7.7 s. The equation allowed for the calculation of t_8/5 cooling times during underwater welding of unalloyed steels using the local dry cavity method based on the variability of the experimental conditions with satisfactory accuracy.