No abstract

Metropolitan railway operators' strategic plans include nowadays actions to reduce energy consumption. The application of ecodriving initiatives in lines equipped with automatic train operation (ATO) systems can provide important savings with low investments. Previous studies carried out under the ATO framework have not considered the main uncertainties in the traffic operation: the train load and delays in the line. This paper proposes a method to design robust and efficient speed profiles to be programmed in the ATO equipment of a metro line. First, an optimal Pareto front for ATO speed profiles that are robust to changes in train load is constructed. There are two objectives: running time and energy consumption. A robust optimization technique and an alternative method based on the conservation of the shape of the speed profiles (pattern robustness) are compared. Both procedures make use of a multi objective particle swarm optimization algorithm. Then, the set of speed profiles to be programmed in the ATO equipment is selected from the robust Pareto front by means of an optimization model. This model is a particle swarm optimization algorithm (PSO) to minimize the total energy consumption considering the statistical information about delays in the line. This procedure has been applied to a case study. The results showed that the pattern robustness is more restrictive and meaningful than the robust optimization technique as it provides information about shapes that are more comfortable for passengers. In addition, the use of statistical information about delays provides additional energy savings between 3% and 14%.Index Terms-Communication-based train control (CBTC), energy saving, multi objective particle swarm optimization (MOPSO), subway systems, train load variations, train operation, uncertainty.

No abstract

No abstract

Digital images often become corrupted by undesirable noise during the process of acquisition, compression, storage, and transmission. Although the kinds of digital noise are varied, current denoising studies focus on denoising only a single and specific kind of noise using a devoted deep-learning model. Lack of generalization is a major limitation of these models. They cannot be extended to filter image noises other than those for which they are designed. This study deals with the design and training of a generalized deep learning denoising model that can remove five different kinds of noise from any digital image: Gaussian noise, salt-and-pepper noise, clipped whites, clipped blacks, and camera shake. The denoising model is constructed on the standard segmentation U-Net architecture and has three variants—U-Net with Group Normalization, Residual U-Net, and Dense U-Net. The combination of adversarial and L1 norm loss function re-produces sharply denoised images and show performance improvement over the standard U-Net, Denoising Convolutional Neural Network (DnCNN), and Wide Interface Network (WIN5RB) denoising models.