Search

The provision of accurate travel time information holds paramount significance for optimizing traffic operations and management. However, existing approaches typically address multiple travel time learning tasks independently, neglecting the correlation of tasks and necessitating the construction of separate models for each task. As the scale of the road network expands, this practice becomes increasingly infeasible. To bridge this gap, this paper presents an adaptive deep multi-task learning model to solve multiple travel time estimation tasks collaboratively from a citywide perspective. The novel model seamlessly integrates geographical, environmental, and spatial-correlated features, capturing the complex non-linear characteristics of travel time in the urban road network through a deep network learning feature representations layer by layer from the input data. Through the paradigm of deep multi-task collaborative learning, the knowledge learned from one task serves as an inductive bias for other tasks, thereby facilitating the learning of related tasks. By modeling the uncertainty of travel time with a probabilistic model, the weights of jointly learned tasks are adaptively fine-tuned in a principled way. Additionally, a novel algorithm is devised to fully leverage the information embedded within the data with partial information. The proposed methodology is evaluated through a case study on the citywide road network in Beijing, China. Empirical results of extensive experiments demonstrate the adeptness of the proposed model in effectively harnessing information transferred from parallel tasks, adaptively fine-tuning the weights of joint learning tasks, and proficiently exploiting data characterized by incomplete labels, therefore leading to superior performance compared to competing methods. The average estimation error of the four jointly learned travel time estimation tasks is about 28.48%, yielding an improvement of over 4% compared to single-task learning models. The results verify the effectiveness and robustness of the proposed model for travel time estimation in the urban road network. [ABSTRACT FROM AUTHOR]

• Accurate risk path forecasting in rail transit systems is a significant and challenging task. • A probabilistic deep learning framework that can process multisource data for risk path forecasting. • The model can effectively handle safety issues such as cascade failures. Failure-related urban rail events can disrupt transit system operations and traffic flow and lead to serious safety problems worldwide. A domino effect can occur if potential cascading events of major failures are not effectively mitigated and controlled. Therefore, accurate risk path forecasting in rail transit systems is a significant and challenging task. Because of the limitations of traditional models in terms of computational power and feature extraction capabilities, this paper proposes a probabilistic deep learning framework that can process multisource data for risk path forecasting in urban rail transit. This paper first proposes a method for constructing a large-scale risk path ground truth dataset when fault events occur. Then, the framework uses a graph-based feature mapping method to model social media, passenger flow, and station failure information. Finally, we proposed a spatiotemporal feature extractor and a dynamic difference weighting loss function to extract features and optimize parameters. We apply real-world data from 2018 to 2019 from the Beijing urban rail transit system for experimental analysis. The analyzed results demonstrate that the proposed model exceeds the baseline models by at least 2.9% in terms of F 1 value, fusing multi-source data exceeds using single-source data by at least 14% in terms of F 1 value and the proposed attention mechanism and dynamic loss function weights can effectively improve the forecasting performance of the model. Furthermore, the results of the different steps ahead demonstrate that the above results are robust. The application of the model can effectively handle safety issues such as cascade failures. [ABSTRACT FROM AUTHOR]

Accurate PM 2.5 forecasting provides a possibility for establishing an early warning system to notify the public and take precautionary measures to prevent negative effects on ambient air quality and public health. Considering strong seasonal variation in meteorological conditions, in this paper, a seasonal stacked autoencoder model combining seasonal analysis and deep feature learning is proposed for forecasting the hourly PM 2.5 concentration, named DL-SSAE model. The original data are firstly decomposed into four seasonal subseries according to the Chinese calendar, and then the Kendall correlation coefficient method is employed to search inherent relationships between PM 2.5 concentrations and meteorological parameters within 1-h ahead for each seasonal time series. The inherent relationships of each seasonal subseries are finally extracted, learned, and modeled by different deep neural networks (stacked autoencoders for regression), and the hourly PM 2.5 forecasts are yielded. The addressed model is tested by the dataset collected from three environmental monitoring stations in Beijing, China. The results demonstrate that the proposed model outperforms all other considered models with/without seasonality consideration in this paper. [ABSTRACT FROM AUTHOR]

Accurate traffic congestion estimation and prediction are critical building blocks for smart trip planning and rerouting decisions in transportation systems. Over the decades, there have been many studies focusing on traffic congestion estimation and prediction with different statistical approaches (e.g., Markov chain) and machine learning models (e.g., clustering, Bayesian networks, and artificial neural networks). However, there is a lack of a unified framework to address the mechanisms of different models and integrate the advantages of different methods through combinations. This paper introduces the FD-Markov-LSTM model, a hybrid interpretable approach that combines the fundamental diagram (FD), Markov chain, and long short-term memory (LSTM). The aim is to estimate and predict traffic states by integrating statistical data in both congested and uncongested scenarios. The FD-Markov-LSTM model leverages the FD to identify hierarchical traffic states and utilizes the Markov process to capture the probabilistic transitions between these states. We employ the LSTM model to further capture the residual time series produced by the Markov chain model (assuming a memoryless property) to enhance the estimation and prediction performance. The proposed model's accuracy in estimating and predicting traffic flow is evaluated using empirical data from three case studies conducted in Beijing and Los Angeles. The results highlight a significant improvement in accuracy compared to classical benchmark models such as the Markov model, ARIMA model, k-Nearest Neighbor model, Random Forest model, and LSTM. Specifically, the FD-Markov-LSTM model achieves reductions of over 39% in mean absolute error, 35% in root mean squared error, and 7.4% in mean absolute percentage error. These results clearly demonstrate that the FD-Markov-LSTM model outperforms the benchmark models, enabling more precise predictions of traffic flow. [ABSTRACT FROM AUTHOR]

• A multitask learning model for short-term inflow prediction of multi-traffic modes is proposed. • The adaptive matrix and attention mechanism are used to extract the spatiotemporal features of multi-traffic modes. • We analyze the interaction mechanism in the multi-traffic modes system. • The research highlights the significance of jointly predicting the inflow of multi-traffic modes. Network-wide short-term inflow prediction is important in efficiently managing the urban transportation system. Nowadays, all kinds of traffic modes gradually become interconnected and form a complex multi-traffic modes system, while extensive studies focus on the single-traffic mode and ignore the correlations among different traffic modes. There exist some challenges for short-term inflow prediction of multi-traffic modes: (1) the interaction mechanism among multi-traffic modes is difficult to learn and few studies explore the mechanism, (2) the data of multi-traffic modes are usually heterogenous due to the different spatial units of different traffic modes, and (3) it is challenging to extract the complex and dynamic features of the multi-traffic modes and most existing methods apply static spatiotemporal correlations among multi-traffic modes, while the genuine correlations among different traffic modes might be missing. To tackle these challenges, this study proposed a multitask-learning-based model called M ulti M ode-former (M2-former) with the encoder-decoder structure for network-wide short-term inflow prediction of the multi-traffic modes system. Specifically, the encoder is designed to learn and capture the complex and dynamic spatiotemporal correlations of multi-traffic modes, and the decoder is designed to extract the features of the target traffic mode and share knowledge among multi-traffic modes. Extensive experiments are conducted based on the real-world multi-traffic modes system data of Beijing, China. Results prove the superiority of the M2-former. In addition, the spatial and temporal information interaction mechanisms among multi-traffic modes are also explored. This paper can provide a reliable method and critical insights for the management and understanding of a multi-traffic modes system. [ABSTRACT FROM AUTHOR]

Forecasting the crowd flows of some densely populated areas (functional areas) at urban level is vital for the city and traffic management and much efforts have been put into this field. However, considering the complex dependencies of crowd flows prediction, the current models generally neglected the multi-scale correlation of spatial view and weakened the influence of external features on crowd flows data. In this paper, we proposed a deep learning framework, called MV-RANet, to deal with them and further improve the prediction performance. Firstly, a double-branch residual attention network was utilized to model the spatial correlation between fine-grained regions and coarse-grained functional areas separately with different branch. And furthermore, some human mobility pattern information were incorporated in MV-RANet to capture the correlation between functional areas in a different view. And as for the impact of external factors, we combined them with crowd flows data by multi-channel mechanism, which can make the external factors and spatio-temporal crowd flows data one-to-one mapping, to refine their influence on the crowd flows. We evaluated our model with two different datasets, a large scale Call Detail Records (CDRs) data and Beijing taxicabs' trajectories data (TaxiBJ) (Zhang et al., 2017). The experimental results demonstrate that the proposed MV-RANet consistently outperforms all the baselines. [ABSTRACT FROM AUTHOR]

• A two-layer deep learning structure based on the LSTM and the GRU was developed to capture the spatio-temporal features of the travel speed of lanes. • Real-world traffic speed data of different lanes in Beijing was used to validate the proposed FDL model. • The proposed FDL model outperforms both the traditional and the state-of-the-art benchmark algorithms, including two deep learning methods. • The predicted speeds of the FDL model are most robust during the 24 h among NN-based benchmark models. Accurate and robust short-term traffic prediction is an important part of advanced traveler information systems. With the development of intelligent navigation and autonomous driving, it is necessary to explore the lane-level predictions of traffic speeds. However, most existing traffic prediction models concentrate on forecasting the traffic flow characteristics of the entire road sections rather than those of certain lanes. This paper proposes a fusion deep learning (FDL) model to predict lane-level traffic speed. First, the entropy-based grey relation analysis is introduced to choose lane sections that are strongly correlated with the lane section to be predicted. Second, a two-layer deep learning framework is established by combining the long short-term memory (LSTM) neural network and the gated recurrent unit (GRU) neural network. Third, the ground-truth data of several lane sections captured by remote traffic microwave sensors (RTMS) on the 2nd Ring Road of Beijing are utilized to examine the FDL model and compare it with several benchmark models. The experimental result indicates that in addition to capturing the fluctuations of traffic speed at the lane level, the FDL model has better performance than the benchmark models in terms of prediction accuracy and stability. [ABSTRACT FROM AUTHOR]

Urban scenes refer to city blocks which are basic units of megacities, they play an important role in citizens' welfare and city management. Remote sensing imagery with largescale coverage and accurate target descriptions, has been regarded as an ideal solution for monitoring the urban environment. However, due to the heterogeneity of remote sensing images, it is difficult to access their geographical content at the object level, let alone understanding urban scenes at the block level. Recently, deep learning-based strategies have been applied to interpret urban scenes with remarkable accuracies. However, the deep neural networks require a substantial number of training samples which are hard to satisfy, especially for high-resolution images. Meanwhile, the crowed-sourced Open Street Map (OSM) data provides rich annotation information about the urban targets but may encounter the problem of insufficient sampling (limited by the places where people can go). As a result, the combination of OSM and remote sensing images for efficient urban scene recognition is urgently needed. In this paper, we present a novel strategy to transfer existing OSM data to high-resolution images for semantic element determination and urban scene understanding. To be specific, the object-based convolutional neural network (OCNN) can be utilized for geographical object detection by feeding it rich semantic elements derived from OSM data. Then, geographical objects are further delineated into their functional labels by integrating points of interest (POIs), which contain rich semantic terms, such as commercial or educational labels. Lastly, the categories of urban scenes are easily acquired from the semantic objects inside. Experimental results indicate that the proposed method has an ability to classify complex urban scenes. The classification accuracies of the Beijing dataset are as high as 91% at the object-level and 88% at the scene level. Additionally, we are probably the first to investigate the object level semantic mapping by incorporating high-resolution images and OSM data of urban areas. Consequently, the method presented is effective in delineating urban scenes that could further boost urban environment monitoring and planning with high-resolution images. [ABSTRACT FROM AUTHOR]

• A deep learning approach to predicting the electric vehicle (EV) charging demand. • A real-world large EV trajectory dataset was used for model training and validation. • The deep learning approach outperformed traditional time series prediction models. • Input data structures, and sample sizes were not influential to prediction accuracy. Short-term prediction of the Electric Vehicle (EV) charging demand is of great importance to the operation of EV fleets and charging stations. This paper develops a Long Short-Term Memory (LSTM) neural network to predict the EV charging demand at the station level for the next few hours (e.g., 1–5 h), using a unique trajectory dataset containing over 76,000 private EVs in Beijing in January 2018. To explore the performance of the LSTM model, we set up four scenarios by 1) comparing LSTM against two typical time series prediction models, i.e., the Auto-Regressive Moving Average model (ARIMA), and the Multiple Layer Perceptron model (MLP), 2) and investigating how different input data structures, sample sizes, and time spans and intervals would influence model accuracy. The results suggest that the LSTM model outperformed the ARIMA, and MLP models, and their MAPE1 values are 6.83 %, 21.58 %, and 18.31 %, respectively. In addition, we find that the time span and interval tend to be more influential to the LSTM model's prediction accuracy than input data structures, and sample sizes. In general, the LSTM model with a shorter time span or interval (e.g., 1 h) would perform better. [ABSTRACT FROM AUTHOR]

Battery swap stations have become an important alternative to general charging posts. Predicting battery swapping demand at the station level would be helpful for real-time operation of stations. This paper first provided insights into battery swapping demand patterns by analyzing a real-world dataset which contained 2,529 battery swapping events collected from 36 battery swap stations in Beijing from 31st July to 20th August 2019. Further, we developed a series of deep learning methods to predict the EV battery swapping demand, particularly considering temporal demand patterns obtained from the dataset. The deep learning models were Long Short-Term Memory, Bidirectional Long Short-Term Memory, Gated Recurrent Units, and Bidirectional Gated Recurrent Units. The results showed that the four deep learning models outperformed typical machine learning methods (e.g., support vector regression). An ablation study indicated that incorporating temporal battery swapping demand patterns into the deep learning methods could greatly improve model performance. [ABSTRACT FROM AUTHOR]

Because renting a house is a low-frequency behavior of people, the housing rental suggestion issue suffers from extremely sparse data. In this paper, we propose to investigate the issue based on e-commerce data, as the two scenarios share the same users, and, more importantly, we can characterize the users in terms of their consumption attitudes and behaviors (that can direct people to rent houses) from the massive online shopping orders, and extract labeled users (users who have both housing and working addresses) from the delivery records. Though e-commerce data can be utilized, developing such an approach is still challenging. This happens because only a tiny part of these users can be used as supervised information and it is non-trivial to fuse the e-commerce data with geographic and traffic data. To this end, we first propose to formulate the task pairwise way to use the labeled users more effectively, the target of which is to evaluate a user's satisfaction degree for a given house. Thereafter, we carefully design features of users and houses from the multi-source data and devise a novel deep network, entitled HouseCritic, to fuse the features and tackle the formulated problem. Notably, HouseCritic introduces meta-learning technology to design a fusion structure that can explicitly and effectively fuse extracted features and accordingly infer a satisfaction degree. Moreover, HouseCritic proposes to leverage unsupervised information, capable of effectively fusing data when the supervised information is insufficient. Finally, we conduct experiments on real-world data collected from Beijing, China. Experimental results demonstrate the advantages of the proposed approach over several baselines. Moreover, a system with the approach is being deployed in an e-commerce company. [ABSTRACT FROM AUTHOR]

Complete and high-quality deformation monitoring data are critical for shield tunnel construction safety and quality. In engineering practices, data missing frequently occurs during instrumentation, adversely impacting further analysis and decision making. Existing imputation methods either ignore the crucial interactions between different parameters during shield tunneling or focus on the global characteristics of deformation data while neglect their local difference. This paper proposes a novel hybrid model, MCCB, combining multi-view matrix completion algorithms, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) algorithms to impute missing deformation values in shield tunnel monitoring data. The performance of the proposed method is verified using bridge deformation data from a practical project in Beijing. Different missing patterns of the bridge deformation data are filled. The experiment results show that the proposed model can effectively learn the various characteristics of the deformation data and outperforms the four selected models and its two sub-models, and can be used to improve the accuracy of the deformation prediction through data imputation. The novelty of this study includes two aspects. First is that the complicated interaction between different parameters and local difference of the data are considered simultaneously. They have not been addressed before by existing studies. Second is that the innovative combination of matrix completion and deep learning algorithm for application in missing deformation values imputation. To our best knowledge, no research on engineering construction has implemented this technique before. [ABSTRACT FROM AUTHOR]

Long short-term memory (LSTM), especially vanilla LSTM (VLSTM), has been widely used in air quality prediction field. However, VLSTM has many more parameters, thereby making training slow and prediction performance unstable. The VLSTM network input data have not been selected for better efficiency. In this paper, we propose an air quality prediction model based on the improved VLSTM with multichannel input and multiroute output (IVLSTM-MCMR). The proposed model includes the IVLSTM and MCMR modules. The proposed IVLSTM module is developed by improving the VLSTM inner structure of VLSTM in order to reduce the number of parameters that help to accelerate the convergence. A new historical information usage approach is further proposed to obtain a stable training process. For the MCMR module, a multichannel data input model (MC) with an improved linear similarity dynamic time warping is introduced to choose the valid data as the input of IVLSTM. A multiroute output model (MR) is designed to integrate the results from MC, in which the results of different target stations with different features are output by different routes. We evaluate the proposed model with the collected data from Beijing, China, and the experimental results show that our model achieves improvements regarding the predication performance. • An air quality prediction model IVLSTM-MCMR is proposed based on deep learning. • The number of parameters in IVLSTM-MCMR is reduced to accelerate the convergence. • The improved linear similarity dynamic time warping is introduced in IVLSTM-MCMR. • The integration of multi-channel data input and multi-route output is designed. [ABSTRACT FROM AUTHOR]

The urban built environment has been extensively studied as a means to reduce car use, but most of the built environment features that have been examined mostly include so-called 'D-variables', namely, density, (functional) diversity, (road network) design, etc. The influence of the three-dimensional street space, which cannot be directly understood from maps, is much less studied. However, characteristics of the street space are considered to influence people's travel behavior according to urban design and walkability theories. A major reason for this gap is the lack of data providing relevant information. This paper aims to fill the gap by harnessing new urban data to measure these characteristics of the urban environment and analyze their impact on travel behavior, as indicated by the amount of car use. Among the many characteristics of street space, we chose the quality of street buildings, continuity of street walls, and height-width ratio of the street valley as three key characteristics to study. Using Beijing as a case study, we found that the three street space characteristics had a statistically significant relationship with residents' car use, mainly by influencing travel mode choice. The continuity of street walls is the most influential factor in reducing car use. • We expanded existing research by examining the influence of the street space on urban car use. • The streetscape is measured using street view images, deep learning models and building footprint data. • The continuity of street walls is the most influential streetscape feature in reducing car use. [ABSTRACT FROM AUTHOR]

[Display omitted] • The measurement errors in long-time series remote sensing images are corrected by statistical and spatio-temporal prediction model. • The ecological degradation of the Miyun Reservoir basin from 1985 to 2021 is mainly caused by anthropogenic factors. • Through statistical theory, the ecological development trend in the experimental area is analyzed, and the conclusion is verified by land use evolution. Miyun Reservoir, located in the Miyun District, Beijing, China, is the largest comprehensive water conservancy project and is an important ecological protection area in the North China region. Changes within the basin are the driving factors affecting the ecosystem in the watershed; therefore, it is important to analyze the changes in the ecological environment of Miyun Reservoir. For the analysis of a long time series of image data remotely sensed by satellite, the outliers caused by atmospheric, lighting, and sensor measurement errors are significant, and it is difficult for traditional algorithms to effectively recover the true image value. To address this, this paper proposes a theoretical model for predicting spatio-temporal variation based on deep learning to identify and correct invalid and anomalous values in extended time series data. This study corrected and analyzed the results of Remote Sensing based Ecological Index inversion of Landsat data of the Miyun Reservoir watershed from 1985 to 2021. The findings and conclusions of this study are important for the analysis of long time series image data from satellite remote sensing and for improving regional ecological evaluation and sustainable development planning. [ABSTRACT FROM AUTHOR]

• Bayesian inference and deep learning is integrated for traffic congestion modeling; • A dynamic Bayesian graph convolutional network (DBGCN) is proposed. • The DBGCN outperforms the state-of-the-art prediction models. • The DBGCN can simulate congestion evolution via dynamic adjacency matrix. • The change of congestion source location leads to different congestion patterns. Congestion, whether recurrent or non-recurrent, propagates through the road network. The process of congestion propagation from a particular road to its neighbors can be regarded as a kind of message passing with a directed relationship. Existing methods have created a solid foundation for characterizing congestion propagation; however, they are either built upon simplified assumptions in traffic flow theory or predefined relationships among road sections, which would lead to downgraded accuracy in practice. This paper proposes a dynamic Bayesian graph convolutional network (DBGCN), which integrates Bayesian inference into a deep learning framework. Therefore, the rules of congestion propagation in the network can be actively learned from the observed data instead of predefining them based on prior knowledge. Experimental results on 971 testbeds in a regional road network in Beijing demonstrate that DBGCN outperforms the state-of-the-art models in inferring the congestion propagation spatiotemporal coverage and reveals variations in congestion propagation patterns according to the road network structure. Furthermore, the proposed model can simulate the congestion propagation process in customized scenarios by learning the latent congestion propagation rules. The results in different scenarios show that the change of congestion source location leads to distinct congestion magnitude, and the propagation of congestion will eventually stop at the road sections with strong shunting effect. [ABSTRACT FROM AUTHOR]