Search

In a high-speed railway (HSR) corridor, railway companies provide diverse train services with different features such as departure time, running speed, and fare. Furthermore, passengers have varying preferences, which influence their travel choices. Focusing on this problem, this paper proposed a novel train timetable optimization model based on departure time and pricing strategy to analyze their impact on passengers’ travel choices with a consistent operating speed. Additionally, an extended time-space-fare network is constructed to visualize this process. The generalized cost of each arc in the network is used as a criterion for evaluating passenger selection. Then, we adapted a passenger flow distribution method based on the user equilibrium (UE) principle to allocate the passenger flow to each train. We also developed a bi-level planning model with passenger allocation as the lower-level model and timetable decision as the upper-level one to optimize collaboratively. Finally, a genetic algorithm integrated with the Frank-Wolfe method is designed to seek the optimal timetable and pricing strategy and a concrete numerical example in Lanzhou-Xi’an HSR is taken to investigate the effectiveness of the proposed model and algorithm.

In China, the transportation capacity of high-speed railways is gradually sufficient to provide services for high-value express freight besides meeting passenger demands. However, similar as the passengers, express freight demands fluctuate and show clear peak and trough periods daily. Therefore, optimizing running numbers of express freight trains on high-speed railway by periods is quite necessary to guarantee the revenues of railway industry and to meet the various requirements of all consignees simultaneously. First, a space-period-pattern three-dimensional network with virtual arcs is built to describe the departure period selection and the stopping or skip-stopping operations. Second, by constructing the arcs' impedance function, the user equilibrium principle is introduced to optimize the express freight flow distribution for each pattern in each period. To elaborate the comprehensive goal of balancing the relationship between profits and the flow distribution, a bi-level programming model is established. The upper model addresses the railway industry's maximum profits, and the lower model addresses the minimum and similar impedance values of the final express freight flow distribution. Finally, through the use of a hybrid algorithm that combines the heuristic genetic algorithm with the Frank-Wolfe algorithm, an experimental case of 10 stations on the Beijing-Xi'an high-speed railway corridor is used to validate the study. The results show that the express freight volume is reasonably distributed to each kind of virtual arc, the impedance value of each pattern is minimized and almost equalized, and the profits of railway industry are maximized by optimizing the number and departure time of trains by each pattern in each period on the basis of meeting all the express freight demands.

Person segmentation in images has various applications, for example, smart home, human-computer interaction, and scene perception for self-driving cars, which are a key feature of the Internet of Things. Due to limitations in performance, such as accuracy and runtime, most traditional methods do not fulfill the practical requirements. Deep learning-based modern segmentation systems become prevalent. Fully convolutional network (FCN), as a classic image semantic segmentation method, directly optimizes the semantic map from the original image in a pixel-wise manner without using pixel-correlations or global object information. In this paper, we propose an efficient end-to-end person segmentation network structure fusing the person detection network with the FCN. The person detection network estimates the region of interest of persons and enforces the segmentation network to focus on the optimization of person segmentation. The loss function of the proposed network considers both the segmentation error and the detection bias error. In addition, the lightweight design of the detection network that optimizes only person bounding-box coordinates enables real-time person detection. The experimental comparison and analysis of several different networks on several datasets show the effectiveness of the proposed fusion strategy. The approach shows a promising practical application potential by fast running time and high segmentation accuracy.

Snowflakes attached to the camera lens can severely affect the visibility of the background scene and compromise the image quality. In this paper, we solve this problem by visually removing snowflakes to convert the snowy image into a clean one. The problem is troublesome; the information about the background of the occluded regions is completely lost for the most part. For removing snowflakes from a single image, we proposed a composition generative adversarial network. Different from the previous generative adversarial networks, our generator network comprises clean background module and a snow mask estimate module. The clean background module aims to generate a clear image from an input snowy image, and snow mask estimate module is used to produce the snow mask in an input image. During the training step, we put forward a composition loss between the input snowy image and composition of the generated clean image and estimated snow mask. We use a dataset named Snow100K2 including indoor and outdoor scenes to train and test the proposed method. The extensive experiments on both synthetic and real-world images show that our network has a good effect and it is superior to the other state-of-the-art methods.

3-D vision plays an important role in intelligent perception of robot, while it requires extra 3-D sensors. Depth estimation from monocular videos provides an alternative mechanism to recover the 3-D information. In this paper, we propose an unsupervised learning framework that uses the perceptual loss for depth estimation. Depth and pose networks are first trained to estimate the depth and the camera motion of the video sequence, respectively. With the estimated depth and pose of the original frame, the adjacent frame can be reconstructed. The pixel-wise differences between the constructed frame and the original frame are used as per-pixel loss. Meanwhile, reconstructed views and original views can be used to extract advanced features from a pre-trained network to define and optimize perceptual loss functions to assess the quality of reconstructions. We combine the respective advantages of these two methods and present an approach of generating a depth map by training the feed-forward network with per-pixel loss function and perceptual loss function. The experimental results show that our method can significantly improve the estimation accuracy of depth map.

Camera node perception capability is one of the crucial issues for visual sensor networks, which belongs to the field of Internet of Things. Multi-object tracking is an important feature in analyzing object trajectories across multiple cameras, thus allowing synthesis of data and security analysis of images in various scenarios. Despite intensive research in the last decades, it remains challenging for tracking systems to perform in real-world situations. We therefore focus on key issues of multi-object state estimation for unconstrained multi-camera systems, e.g., data fusion of multiple sensors and data association. Unlike previous work that rely on camera network topology inference, we construct a graph from 2-D observations of all camera pairs without any assumption of network configuration. We apply the shortest path algorithm to the graph to find fused 3-D observation groups. Our approach is thus able to reject false positive reconstructions automatically, and also minimize computational complexity to guarantee feasible data fusion. Particle filters are applied as the 3-D tracker to form tracklets that integrate local features. These tracklets are modeled by a graph and linked into full tracks incorporating global spatial-temporal features. Experiments on the realworld PETS2009 S2/L1 sequence show the accuracy of our approach. Analyses of the different components of our approach provide meaningful insights for object tracking using multiple cameras. Evidence is provided for selecting the best view for a visual sensor network.

Human action monitoring can be advantageous to remotely monitor the status of patients or elderly person for intelligent healthcare. Human action recognition enables efficient and accurate monitoring of human behaviors, which can exhibit multifaceted complexity attributed to disparities in viewpoints, personality, resolution and motion speed of individuals, etc. The spatial-temporal information plays an important role in the human action recognition. In this paper, we proposed a novel deep learning architecture named as recurrent 3D convolutional neural network (R3D) to extract effective and discriminative spatial-temporal features to be used for action recognition, which enables the capturing of long-range temporal information by aggregating the 3D convolutional network entries to serve as an input to the LSTM (Long Short-Term Memory) architecture. The 3D convolutional network and LSTM are two effective methods for extracting the temporal information. The proposed R3D network integrated these two methods by sharing a shared 3D convolutional network in sliding windows on video streaming to capturing short-term spatial-temporal features into the LSTM. The output features of LSTM encapsulate the longrange spatial-temporal information representing high-level abstraction of the human actions. The proposed algorithm is compared to traditional and the-state-of-the-art and deep learning algorithms. The experimental results demonstrated the effectiveness of the proposed system, which can be used as smart monitoring for remote healthcare.