Short-term statistics of ice loads on ship bow frames in floe ice fields: Full-scale measurements in the Antarctic ocean

Ship operation and ice loading in floe ice fields have received considerable interest during recent years. There have been several numerical simulators developed by different institutes which can simulate ship navigation through floe ice fields and estimate ship performance and local ice loads. However, public data obtained from full-scale measurement covering comprehensively ship performance and ice loads under various ice thicknesses, concentrations and floe sizes are rare. The 2018/19 Antarctic voyage of the Polar Supply and Research Vessel (PSRV) S.A. Agulhas II gathered considerable data of the ship in floe ice fields under various thicknesses, concentrations, and floe sizes. The aim of this paper is to carry out statistical analysis to seek suitable probability distributions which adequately fit the measured ice load and therefore suitable to be used as parent distributions for long-term estimation. For this aim, three categories of probability distributions, namely standard distributions, truncated distributions and mixture distributions are tested. It is found that truncated distributions can fit the load data better than standard distributions bounded at the threshold. In addition, mixture distributions are shown to have promising features, which fit the data well and are able to separate distribution components. Subsequentially, the well-performed distributions are used as parent distributions to make long-term load estimations. The estimation results demonstrate that long-term estimations are sensitive to the selection of parent distribution, which addresses the importance of finding correct distribution to model short-term ice loads. The data of ten selected cases will be published for the use of other researchers.     

机动目标跟踪典型算法评述

机动目标跟踪算法可以划分为两种：极大似然法和贝叶斯估计法，前者通过对目标量测序列的分析，采用与当前目标运动模式匹配度最高（极大似然性）的模型（包括系统模型和观测模型）对目标状态进行跟踪和估计；后者则首先假定一组模型，再根据量测序列进一步估计出其中每个模型的匹配概率，最后采用贝叶斯公式对全部或部分模型估计进行综合。在此框架下，论文对其中的一些典型算法进行总结和分析。     

基于贝叶斯网络的辐射源威胁等级评估算法

贝叶斯网络是近年来人工智能领域最重要的研究成果之一，在具有不确定性的推理和决策问题中得到了广泛的应用。以贝叶斯网络及其推理机制为基础，主要研究了贝叶斯网络在辐射源威胁等级评估方面的应用。选取了部分影响因素，建立了基于贝叶斯网络的辐射源威胁等级评估模型。最后通过一个计算机仿真实例验证了算法具有连续性和累积性，能够对辐射源威胁等级做出较准确评估。     

基于贝叶斯网络的声呐系统战场损伤评估及修复模型

为了解决战场环境下的声呐系统损伤评估及修复（SSBDAR）问题,本文研究了战场损伤评估及修复模型。针对舰艇声呐系统战场损伤评估的因果推理特性,建立了基于贝叶斯网络的战场损伤评估模型,提出了基于贝叶斯网络的评估流程,给出了基于贝叶斯网络的评估决断算法。该模型利用贝叶斯网络对各评估修复节点进行分析,能够解决损伤信息不确定性,而且充分利用舰艇声呐系统各节点的损伤信息,提高了模型评估及修复效率,是解决舰艇声呐系统战场损伤评估及修复问题的一个有效途径。     

安防和乘客异动在途监测系统设计

针对城市轨道交通视频监控的重要性与传统的视频监控的局限性,提出利用前端图像处理模块实现动态背景建模、目标检测和跟踪,对正常的群体行为进行建模、识别乘客异动。设计采用贝叶斯分类器将目标和背景进行分类,实现目标检测,并利用乘客特征的先验知识制成模板,将运动区域与目标模型匹配实现跟踪的过程,并将乘客行为抽象成为时空变化的数据分类问题。本系统的设计旨在保证地铁列车的安全运行,建设与轨道交通相适应的车载视频监控系统。     

基于贝叶斯网络的全自动运行系统弹性评估

结合列车全自动运行系统的弹性评估,提供一种基于贝叶斯网络模型的评估方法。根据列车故障前后的 性能变化曲线,总结系统弹性的 4 种能力,探究影响这些能力的相关因素,为影响因素的定量选取合适的评价指 标,作为贝叶斯网络的最底层节点。对指标制定打分表,获取专家的打分结果作为指标层的先验概率输入,同时 选择恰当的隶属度公式作为条件概率的计算方法,获得贝叶斯网络的条件概率表,建立列车全自动运行系统的贝 叶斯网络模型,定量评估系统弹性,同时对目标节点进行敏感度分析,提出提升弹性的改进建议。     

神朔铁路钢轨折断风险评估模型研究

钢轨折断是一类严重的线路故障,研究钢轨折断风险评估对保障线路安全具有重要作用。基于模糊推理方法构建钢轨折断风险评估模型,该模型利用设备台账数据、钢轨状态检测数据和维修数据等相关的生产数据,识别钢轨折断致灾因子,量化评定致灾因子状态,建立模糊推理规则库,利用Mamdani模糊推理算法计算钢轨折断风险事件发生的可能性。最后采用神朔铁路神木北至黄羊城区间2013～2015年共3年的实际生产数据对模型的有效性进行验证,结果表明:所建模型可以较好地评估神朔铁路钢轨折断风险事件发生的可能性,对钢轨折断风险管理具有重要指导意义。     

基于FTA-BN的综合交通枢纽中地铁车站安全风险评价

以综合交通枢纽中的地铁车站安全为研究目标,综合考虑综合交通枢纽中特殊的客流组成和乘客特征,以踩踏、火灾、水灾、公共卫生和大面积滞留5类易发风险事故作为研究对象,基于FTA-BN方法对其影响因素进行分析,识别其风险因素,建立相应的事故树模型,转化为贝叶斯网络模型进行风险评价;引入三角模糊数处理专家自然语言,得到贝叶斯网络中的先验概率和条件概率分布,然后通过贝叶斯网络模型进行网络推理计算和敏感性分析,找出地铁车站中的薄弱部分,制定相应的风险管控措施,从而提高枢纽中地铁车站对于紧急事件的应对能力。以广州南站地铁车站为例进行快速评价,结果表明：广州南站在公共卫生安全方面存在一定危险,发生概率为42.98%,且较易发生大面积滞留事件,可能性为30.40%。     

Analysis of factors that may be essential in the decision to fly on fully autonomous passenger airliners

This research investigates factors that influence opinion in the decision to fly on fully autonomous passenger airliners primarily from the perspective of aviation and technology professionals. Bayesian statistical inference and a two‐level fractional factorial survey are used to sample passengers' views on fully autonomous airliners. Eight trust, safety, and cost factors are incorporated into a vignette set in the future. Factors include automation levels, safety records, liability guarantees, airline integrity, and service disruptions. Dependent variables exist in five post‐vignette questions and essentially ask “Would you” or “Would you not” be willing to fly on a fully autonomous airliner? Sixteen versions of the vignette, each with unique trust, safety, and cost levels, present varying (unknown) degrees of influence to the survey respondents. For every demographic, the research shows a 99% statistically significant difference between the “prior” and “posterior” sampled population proportions willing to fly. The most significant positive influence involves integrity characteristics of the airline, while the most negative influence relates to life insurance liability guarantees. Research from 2003 suggested that this mode of travel would be acceptable to only 10.5% of respondents. When the 2003 research is used as a Bayesian prior probability, the resulting posterior probability for the demographics sampled can be modeled as a beta distribution, indicating 95% probability that the sampled proportion of the population willing to fly is between 33.2% and 36.4%. After adjusting for age and profession demographics to match the US population, the 95% probability bounds on the proportion willing to fly are 31.35% and 34.15%. Copyright © 2015 John Wiley & Sons, Ltd.     

Recursive estimation of average vehicle time headway using single inductive loop detector data

Vehicle time headway is an important traffic parameter. It affects roadway safety, capacity, and level of service. Single inductive loop detectors are widely deployed in road networks, supplying a wealth of information on the current status of traffic flow. In this paper, we perform Bayesian analysis to online estimate average vehicle time headway using the data collected from a single inductive loop detector. We consider three different scenarios, i.e. light, congested, and disturbed traffic conditions, and have developed a set of unified recursive estimation equations that can be applied to all three scenarios. The computational overhead of updating the estimate is kept to a minimum. The developed recursive method provides an efficient way for the online monitoring of roadway safety and level of service. The method is illustrated using a simulation study and real traffic data.