基于深度学习的驾驶员状态识别

提出了基于驾驶员脸部及周围信息的驾驶员状态检测方法。文章通过实车摄像头采集了驾驶员驾驶状态视频数据,利用Dlib和OpenCV库对采集的驾驶员图像进行脸部检测,基于驾驶员脸部数据建立了深度学习数据集,然后基于该数据集设计了一种卷积神经网络模型FaceNet,利用PyTorch深度学习框架在数据集上对模型进行训练,最终得到了有较高准确率的驾驶员状态检测模型,其可识别抽烟、睡觉、左手打电话和右手打电话四种驾驶员状态。     

Short-term speed predictions exploiting big data on large urban road networks

Big data from floating cars supply a frequent, ubiquitous sampling of traffic conditions on the road network and provide great opportunities for enhanced short-term traffic predictions based on real-time information on the whole network. Two network-based machine learning models, a Bayesian network and a neural network, are formulated with a double star framework that reflects time and space correlation among traffic variables and because of its modular structure is suitable for an automatic implementation on large road networks. Among different mono-dimensional time-series models, a seasonal autoregressive moving average model (SARMA) is selected for comparison. The time-series model is also used in a hybrid modeling framework to provide the Bayesian network with an a priori estimation of the predicted speed, which is then corrected exploiting the information collected on other links. A large floating car data set on a sub-area of the road network of Rome is used for validation. To account for the variable accuracy of the speed estimated from floating car data, a new error indicator is introduced that relates accuracy of prediction to accuracy of measure. Validation results highlighted that the spatial architecture of the Bayesian network is advantageous in standard conditions, where a priori knowledge is more significant, while mono-dimensional time series revealed to be more valuable in the few cases of non-recurrent congestion conditions observed in the data set. The results obtained suggested introducing a supervisor framework that selects the most suitable prediction depending on the detected traffic regimes.     

Airline network choice and market coverage under high-speed rail competition

While the existing literature has focused on the short-term impacts, this paper investigates the long-term impacts of high-speed rail (HSR) competition on airlines. An analytical model is developed to study how an airline may change its network and market coverage when facing HSR competition on trunk routes. We show that prior to HSR competition, an airline is more likely to adopt a fully-connected network and cover fewer fringe markets if the trunk market is large. Under HSR competition, the airline will, for a given network structure, have a greater incentive to cover more fringe (regional or foreign) markets if the trunk market is large, or the airline network is close to hub-and-spoke. Further, the airline will, for any given market coverage, move towards a hub-and-spoke network when the trunk market is large, or the number of fringe markets covered by the airline network is large. Both effects are more prominent when the decreasing rate of airline density economies is large. We further show that HSR competition can induce the airline to adopt network structure and market coverage that are closer to the socially optimal ones, thereby suggesting a new source of welfare gain from HSR based on its long-term impacts on airlines. Implications for operators, policy makers and specific countries (such as China) are also discussed.     

Personalized real-time traffic information provision: Agent-based optimization model and solution framework

The advancement of information and communication technology allows the use of more sophisticated information provision strategies for real-time congested traffic management in a congested network. This paper proposes an agent-based optimization modeling framework to provide personalized traffic information for heterogeneous travelers. Based on a space–time network, a time-dependent link flow based integer programming model is first formulated to optimize various information strategies, including elements of where and when to provide the information, to whom the information is given, and what alternative route information should be suggested. The analytical model can be solved efficiently using off-the-shelf commercial solvers for small-scale network. A Lagrangian Relaxation-based heuristic solution approach is developed for medium to large networks via the use of a mesoscopic dynamic traffic simulator.     

基于文化粒子群网络优化算法的医学图像研究

文章提出了基于文化粒子群的网络优化算法的医学图像方法，通过在50例唇裂矫正术中运用BP神经网络进行学习．对唇部进行解剖观察和美学分析，在矫正修复手术中不断学习和调整来接近期望输出，并通过实验仿真及客观的图像提出了评估准则。实验结果表明，在美学研究基础上的基于文化粒子群BP神经网络的医学图像研究，适用于唇裂的矫正修复。     

Real time detection of driver attention: Emerging solutions based on robust iconic classifiers and dictionary of poses

Real time monitoring of driver attention by computer vision techniques is a key issue in the development of advanced driver assistance systems. While past work mostly focused on structured feature-based approaches, characterized by high computational requirements, emerging technologies based on iconic classifiers recently proved to be good candidates for the implementation of accurate and real-time solutions, characterized by simplicity and automatic fast training stages.In this work the combined use of binary classifiers and iconic data reduction, based on Sanger neural networks, is proposed, detailing critical aspects related to the application of this approach to the specific problem of driving assistance. In particular it is investigated the possibility of a simplified learning stage, based on a small dictionary of poses, that makes the system almost independent from the actual user.On-board experiments demonstrate the effectiveness of the approach, even in case of noise and adverse light conditions. Moreover the system proved unexpected robustness to various categories of users, including people with beard and eyeglasses. Temporal integration of classification results, together with a partial distinction among visual distraction and fatigue effects, make the proposed technology an excellent candidate for the exploration of adaptive and user-centered applications in the automotive field.     

Continuous approximation for skip-stop operation in rail transit

Operating speed of a transit corridor is a key characteristic and has many consequences on its performance. It is generally accepted that an increased operating speed for a given fleet leads to reduced operating costs (per kilometer), travel and waiting times (three changes that can be computed precisely), an improved comfort and level of service, which can attract new passengers who are diverted from automobile (items harder to estimate precisely). That is why several operation schemes which aim to increase the operating speed are studied in the literature, such as deadheading, express services, and stop skipping.A novel category of solutions to this problem for one-way single-track rail transit is to perform accelerated transit operations with fixed stopping schedules. The concept is quite simple: as the time required for stopping at each station is an important part of travel time, reducing it would be a great achievement. Particular operations that take advantage of this idea already exist. This paper focuses on one of them: the skip-stop operation for rail transit lines using a single one-way track. It consists in defining three types of stations: AB stations where all the trains stop, and A and B stations where only half of the trains stop (stations type A and B are allocated interchangeably). This mode of operation is already described in the literature (Vuchic, 1973, Vuchic, 1976, Vuchic, 2005) and has been successfully implemented in the Metro system of Santiago, Chile.This work tackles the problem with a continuous approximation approach. The problem is described with a set of geographically dependent continuous parameters like the density of stations for a given line. Cost functions are built for a traditional (all-stop) operation and for skip-stop operation as described above. A simple example is presented to support this discussion. Finally, a discussion about the type of scenarios in which skip-stop operations are more beneficial is presented.     

Network equilibrium models with battery electric vehicles

The limited driving ranges, the scarcity of recharging stations and potentially long battery recharging or swapping time inevitably affect route choices of drivers of battery electric vehicles (BEVs). When traveling between their origins and destinations, this paper assumes that BEV drivers select routes and decide battery recharging plans to minimize their trip times or costs while making sure to complete their trips without running out of charge. With different considerations of flow dependency of energy consumption of BEVs and recharging time, three mathematical models are formulated to describe the resulting network equilibrium flow distributions on regional or metropolitan road networks. Solution algorithms are proposed to solve these models efficiently. Numerical examples are presented to demonstrate the models and solution algorithms.     

基于BP神经网络的二手车价格评估影响因素研究

为了高效客观准确评估二手车价格,以影响二手车价格评估的因素为研究对象,采用文献法分析筛选出二手车评估价格的11个影响因素并建立了BP神经网络二手车价格评估模型。通过BP神经网络二手车价格评估模型的计算结果显示模型预测价格与实际价格相关系数达到0.96,根据所建模型的连接权值得出了二手车价格评估影响权重值较大的7个因素。最后将影响二手车价格评估的7个主要因素作为输入重构了价格评估模型并重新计算二手车价格得出BP-11模型与BP-7模型计算结果基本一致且Pearson相关性度到0.83。因此本文的研究结果表明:二手车价格评估主要受综合油耗、车辆售后满意度、车龄、车辆可靠性、舒适性、外观、当前里程数的影响,BP神经网络二手车价格评估模型可由这7个因素描述。     

VPN技术在校园网中的应用与实例

虚拟专用网络(VPN)是实现在公网上安全传输私有网络信息的一种技术。VPN技术在校园网中的应用,提高了校园网的可管理性、灵活性和安全性。文章就虚拟专用网(VPN)技术做了详细介绍,并提供了一个简单的VPN组建实例。