A generalized partite-graph method for transportation data association

There are many problems in transportation which involve reconstructing the associations between different entities. For example, data points related to a vehicle from different sensors could be matched to reconstruct the trajectories of vehicles. Or, in population synthesis for microsimulation, lists of persons, dwellings, and vehicles could be generated individually from source data and then matched into synthetic households. There are numerous other examples. The unifying theme is a desire to construct realistic unit-level associations from aggregate or anonymized data. The problem demands a method that is behaviorally consistent and operationally efficient to handle large datasets. We adapt concepts from graph theory to formulate this class of problems as a k-partite graph. This approach is generic and can incorporate expectations of behavior in the form of edge weights. A Dijkstra algorithm based solution is proposed for a subset of k-partite graphs which permits a direct comparison with pair-wise matching and applied to a case study of bicycle tracklets. We then propose an iterative improvement algorithm as a generic method and apply it to a complete k-partite graph in a population synthesis case study. The first case study shows that the k-partite algorithm outperforms the previously used pair-wise matching algorithms. The second case study demonstrates the generality of the proposed algorithm to all k-partite graphs and shows that the generic method is fast and scalable to large problems. As a whole, this paper aims to show that k-partite methods are behaviorally consistent, efficient, and potentially applicable to a wide variety of transportation data association problems.     

Real-time identification of probe vehicle trajectories in the mixed traffic corridor

This paper proposes three enhanced semi-supervised clustering algorithms, namely the Constrained-K-Means (CKM), the Seeded-K-Means (SKM), and the Semi-Supervised Fuzzy c-Means (SFCM), to identify probe vehicle trajectories in the mixed traffic corridor. The proposed algorithms are able to take advantage of the strengthens of topological relation judgment and the semi-supervised learning technique by optimizing the selection of pre-labeling samples and initial clustering centers of the original semi-supervised learning technique based on horizontal Global Positioning System data. The proposed algorithms were validated and evaluated based on the probe vehicle data collected at two mixed corridors on Shanghai’s urban expressways. Results indicate that the enhanced SFCM algorithm could achieve the best performance in terms of clustering purity and Normalized Mutual Information, followed by the CKM algorithm and the SKM algorithm. It may reach a nearly 100% clustering purity for the uncongested conditions and a clustering purity greater than 80% for the congested conditions. Meanwhile, it could improve clustering purity averagely by 21% and 14% for the congested conditions and 6.5% and 6% for the uncongested conditions, as compared with the traditional K-Means algorithm and the basic SFCM. The proposed algorithms can be applied for both on-line and off-line purposes, without the need of historical data. Clustering accuracies under different traffic conditions and possible improvements with the use of historical data are also discussed.     

智能船舶航线优化方法综述

为提升船舶轨迹异常检测的精度和效率,解决传统异常检测方法存在的特征表征能力有限、补偿精度不足、容易出现梯度消失、过拟合等问题,提出一种基于Transformer_LSTM编解码器模型的无监督船舶轨迹异常检测方法。

该方法基于编码器解码器架构,由Transformer_LSTM模块替代传统神经网络实现轨迹特征提取和轨迹重构;将Transformer嵌入LSTM的递归机制,结合循环单元和注意力机制,利用自注意力和交叉注意力实现对循环单元状态向量的计算,实现对长序列模型的有效构建;通过最小化重构输出和原始输入之间的差异,使模型学习一般轨迹的特征和运动模式,将重构误差大于异常阈值的轨迹判定为异常轨迹。

采用2021年1月的船舶AIS数据进行实验,结果表明,模型在准确率、精确率以及召回率上相较于LOF,DBSCAN,VAE,LSTM等经典模型有着明显提升;F₁分数相较于VAE_LSTM模型提升约8.11%。

该方法的异常检测性能在各项指标上显著优于传统算法,可有效、可靠地运用于海上船舶轨迹异常检测。

     

基于轨迹线性化控制方法的全驱动自主式水下航行器轨迹跟踪控制

  目的  针对欠驱动水面船舶在轨迹跟踪控制中存在的动态不确定和未知扰动问题,设计一种自适应神经网络有限时间轨迹跟踪控制方案。  方法  利用运动学虚拟控制律变换和有界限制的方法进行欠驱动变化。在Backstepping的框架下,利用神经网络重构未知动态,并设计自适应律逼近未知扰动的上界。通过Lyapunov直接法提供严格的理论分析,以证明闭环系统所有信号都是有界的,并使跟踪误差收敛至有界的区间。  结果  仿真结果表明,所提控制方案能够使欠驱动船舶在有限的时间内跟踪上期望的轨迹,且相比传统控制方案,系统误差的收敛速度更快,误差的上、下界也更小,在面对外界未知的时变干扰时还具有良好的鲁棒性。  结论  所做研究可为船舶的轨迹跟踪控制提供有效参考,具有实际的工程意义。     

基于双视角学习原理的交通视频车辆事件鲁棒检测

提出了1种基于双视角学习原理的高速公路交通视频车辆事件鲁棒检测算法.针对道路交通结构化特点提出了分车道外极面图(Epipolar Plane Image,简称 EPI),以此反映交通断面车流整体特征.基于双视角学习原理,融合现有广泛应用的反映车辆独立行为的行驶轨迹特征,实现高速公路车辆事件鲁棒检测.针对多种典型车辆事件(包括交通拥堵,车辆逆行,车辆违规停车,交通事故等),本文算法总体检测率为94.09%,误检率为4.51%,漏检率为1.40%,其性能与传统单视角方法比较有较大的提高.     

纵桁穿过边缘开孔舱壁对环肋圆柱壳水下声辐射特性的影响

为研究纵桁穿过边缘开孔舱壁对环肋圆柱壳水下声辐射特性的影响,采用结构有限元耦合流体边界元方法,通过FORTRAN和DMAP混合编程处理流体附加质量和附加阻尼,计算在不考虑开孔加强结构的条件下,不同舱壁边缘开孔数量、半径及分布角度的圆柱壳表面均方法向速度级与辐射声功率级频响曲线。结果表明:舱壁边缘开孔分布角度对环肋圆柱壳湿表面均方法向速度级和辐射声功率级的分贝值影响很小,但对两声学衡量指标的峰值数影响明显;边缘开孔数量、半径对两声学衡量指标的分贝值影响明显,且对峰值数影响显著。因此,在设计潜艇舱壁结构时,需合理设置边缘开孔参数。     

流致噪声机理及预报方法研究综述

从自由湍流噪声、壁面湍流噪声、转子噪声和空腔流动4 个方面对流致噪声机理及预报方法进行综述。对目前工程应用中的3个主要流致噪声预报方法,即Lighthill声比拟理论、Kirchhoff方法和涡声理论的基本原理及适用性进行详细讨论,并对流致噪声数值模拟方法进行总结。其中,Lighthill声比拟理论属噪声源先验理论,虽方便应用但不能描述声流相互作用基础问题;Kirchhoff方法在运用的过程中虽不需要确切获知源的属性,但声源区的计算精度很重要;涡声理论在声流相互作用等领域有着良好的研究前景。自由湍流噪声以四极子雷诺应力源为主,存在如螺旋桨等固壁边界时则会产生偶极子源,在低马赫数流动中是更为有效的声源。     

基于不同模型的正交加筋板低频振动和声辐射分析

正交加筋薄板在船舶结构中应用十分广泛。从简化加筋板模型的角度出发,基于李兹法能量泛函变分原理推导加筋板的能量求解;并针对双向加筋板基于正交异性板理论和相当板厚理论提出2种简化方法,对3种模型模拟的加筋板推导其固有频率的解析公式。计算表明：3种正交加筋板模型在求取低阶固有频率时均具有良好的精度。在此基础上研究了基于3种模型的加筋板在空气中和水中的声辐射特性,为加筋板组合结构的振动和低频声辐射问题提出了新的简化思路。     

有限水域试验场声传播特性的仿真计算与试验

在有限水域试验场中进行大型复杂结构的水下声振试验,声信号测量和数据处理是关键。天然水域边界条件未知,水文环境复杂,仿真难度高,增加了数据处理难度。采用Matlab和VB混合编程软件AcTUP,分析试验场“沉积层和水位”的选取原则,得到与试验值较为吻合的仿真模型。通过仿真计算,得到固定深度声源的“无因次量纲”参数、固定水深的声源位置及水面粗糙度对试验场声传播损失的影响规律,并基于射线声学理论对试验值和仿真值的偏差进行非相干分析,得到了有限水域试验场的基本声传播特性。     

无人艇工程控制课程实践教学平台开发

针对在教学环节中存在的问题,利用计算机技术和工业软件开发无人艇工程控制课程实践教学平台。该平台可进行无人水面艇（Unmanned Surface Vessel,USV）建模及运动轨迹控制。针对USV物理参数的不确定性,根据运动学和动力学模型设计控制器,分析控制器参数对USV运动性能的影响规律。无人艇工程控制课程实践教学平台的建设应用可提升学生对USV的运动操控、故障判断及维护保养能力。