Destination prediction is an essential problem for many location-based applications and services. Although previous works partly solved the sparsity of GPS location data by methods such as discretization and embedding, the problem of properly extracting and utilizing geographical information of trajectories is still unsolved. The paper proposes the TransGTE model, a Transformer-based framework with a novel geographical embedding and fusion mechanism, to adaptively extract and fuse geographical features with trajectories’ sequential patterns. TransGTE uses the Graph Convolutional Network (GCN) and Transformer to extract geographical and sequential features and adopts a dynamic gating mechanism to control the weights of sequential and geographical information adaptively. We perform extensive experiments on four taxi trajectory real-world datasets from Porto, Chengdu, Shenzhen and San Francisco, where the TransGTE averagely outperform the best benchmark models by 4.24%, 2.87%, 5.91% and 4.11% in terms of the Mean Haversine Distance Error. The ablation study validates the effectiveness of the proposed trajectory location representation and dynamic gating mechanism modules used to embed taxi GPS trajectories. Finally, we compare the proposed trajectory embedding with the commonly used transformer-based model, and it highlights the effectiveness of the proposed embedding approach in representing geographical similarities between trajectories. The code for this paper is available at: https://github.com/qzl408011458/TransGTE.

Short-term route prediction on road networks allows us to anticipate the future trajectories of road users, enabling various applications ranging from dynamic traffic control to personalized navigation. Despite recent advances in this area, existing methods focus primarily on learning sequential transition patterns, neglecting the inherent spatial relations in road networks that can affect human routing decisions. To fill this gap, this paper introduces RouteKG, a novel Knowledge Graph-based framework for route prediction. Specifically, we construct a Knowledge Graph on the road network to encode spatial relations, especially moving directions that are crucial for human navigation. Moreover, an n-ary tree-based algorithm is introduced to efficiently generate top-K routes in batch mode, enhancing computational efficiency. To further optimize prediction performance, a rank refinement module is incorporated to fine-tune candidate route rankings. The model performance is evaluated using two real-world vehicle trajectory datasets from two Chinese cities under various practical scenarios. The results demonstrate a significant improvement in accuracy over the baseline methods. We further validate the proposed method by utilizing the pre-trained model as a simulator for real-time traffic flow estimation at the link level. RouteKG has great potential to transform vehicle navigation, traffic management, and a variety of intelligent transportation tasks, playing a crucial role in advancing the core foundation of intelligent and connected urban systems. The source codes of RouteKG are available athttps://github.com/YihongT/RouteKG

Urban rail transit systems in many cities are experiencing crowding during peak periods due to rapid population growth. Incentive-based demand management strategies aim to better utilize the available capacity by shifting peak travel to off-peak periods. Various deployments have demonstrated the crowding-reduction potential of incentives in reducing crowding but they have also shown that such strategies are inefficient with many passengers receiving the incentives but relatively few contributing to crowding reduction. This paper proposes a reverse auction-based, individualized incentive strategy to encourage individual passengers to switch travel from peak to off-peak periods. The proposed approach is individualized, participatory, and explicitly accounts for individual characteristics and the potential contribution of their behavior changes to the system. Extensive experiments are conducted to demonstrate the approach using AFC data from Hong Kong’s urban rail network. The results indicate that auction-based individualized incentives can enhance the system efficiency by strategically selecting passengers as winners whose behavioral changes contribute to the system performance. It also highlights the importance of correcting the information bias of perceived travel behavior between bidders and the population when operators select bid winners.

Eye contact is essential in transmitting information and intention in the wild environment (e.g., urban streets or parking lots) with mixed vehicles and pedestrians. Compared with the vision image data, the human skeleton data are deemed to be robust to unconstrained surroundings and illumination. However, the skeleton graph-based approaches are mainly used for the action recognition. It is challenging to directly apply them to the eye detection task, which is momentary and dynamic given the complex wild environment. This paper proposes a Bi-stream Spatial Attention Graph Convolution Network (SA-BiGCN) for eye contact detection in the wild. We design a directed, nose-centric skeleton graph to capture relevant and hierarchical information and their interactions. We also propose a Bi-stream graph convolution network model with spatial attention to dynamically extract and fuse skeleton joints and bones information. The model was validated by comparing with state-of-art models on three large-scale public datasets, including JAAD, PIE, and LOOK. The results highlight the accuracy and generalization performance of the proposed SA-BiGCN model in detecting the eye contact in the wild environment. The ablation analysis validates the importance of the skeleton graph design, the spatial attention mechanism in the feature fusion process, as well as the model robustness against noisy skeleton data in terms of part occlusions, block occlusions, random occlusions, and random deviations.

Accurately and timely predicting pedestrian crossing intentions in real-time is critical for operating intelligent vehicles on roads. Although existing models achieve promising accuracy using complex models and video image data, they are constrained for real-time practical use given the high model complexity, time-consuming data preprocessing, and low-quality image data in the wild. To address these, the paper proposes a Spatial-Temporal Attention Graph Convolution Network model for fast pedestrian crossing intention prediction (PedAST-GCN). It uses a lightweight GCN model as the backbone network with simple but robust graph representations of pedestrian crossing intention modality features, including pedestrian pose, bounding box, and vehicle speeds. The model is validated by comparing it with state-of-the-art models on two large-scale public datasets (JAAD and PIE). The results highlight the better performance of the PedAST-GCN model for pedestrian crossing intention prediction in terms of accuracy and computation times. The ablation analysis confirms the value of the backbone layer and graph design, the designed modality features, the effectiveness of attention mechanisms in capturing long-term dependencies (spatial-temporal attention) and fusing heterogeneous features (modality attention), and the robust performance across various observation lengths and in the presence of noisy data.

This paper studies the estimation of dynamic route choice behavior of drivers with incomplete fixed location-based sensor data, such as radio frequency identification (RFID) data. Unlike global positioning system (GPS) data providing continuous vehicle trajectories, the location-based RFID sensors record vehicles only when they pass by but may not record all vehicles. These bring challenges for route choice modeling since empty sensor observations will also influence the likelihood of routes that do not cross these sensors. Also, it requires the essential integration of dynamic traffic conditions into the modeling process as observation paths may share the same sensor detection sequence but exhibit different travel times. To address these challenges, the paper proposes a dynamic recursive logit model to estimate vehicle route choices with RFID data, enabling the characterization of the likelihood function of sensor observation paths without the need for path enumeration between consecutive detections. Also, we develop a probabilistic dynamic link utilization estimation method to infer the actual path of each vehicle from the available sensor observations. It serves as a validation process to ensure that the route choice behavior can comprehensively reflect traffic flow dynamics. The proposed methods are evaluated using both a simulated dataset on the Sioux Falls network and a collection of real-world RFID data in Chongqing, China. The simulation results show that the proposed method can recover true choice parameters and perform significantly better compared with static models. The real-world experimental results highlight its efficacy in aggregated link flow prediction and individual trajectory reconstruction.

Signal phase and timing (SPAT) is critical to the operation of signalized intersections. In practice, data-sharing barrier makes it challenging to acquire SPAT data, impeding traffic management applications. Existing SPAT estimation methods mainly focus on partial estimation for selected time periods, and use data from multiple days, with a potentially problematic assumption that the same signal timing plan is applied across all the days. To address these challenges, this study proposes a method to estimate SPAT information using trajectory data from radar vision integrated camera (RVIC) sensors. The method is validated against real-world traffic light state data in China. The results show that the method can make an accurate estimation of time-of-day (TOD) division, cycle length, phasing scheme, and phase duration. It outperforms previous studies in cases with a limited number of observations. Further, with a sensitivity test, we derive the minimum amount of data required for the proposed method, which is quantified explicitly using the capture rate of the actual green-start times of each movement. The method estimates SPAT information daily, supporting signal evaluation and optimization functionalities. It can also drive Artificial Intelligence - Internet of Things applications, for example, providing red or green light countdown for drivers and extending infrastructure coverage for automated and connected vehicles.

Abstract Understanding human mobility in urban areas is important for transportation, from planning to operations and online control. This paper proposes the concept of user-station attention, which describes the user’s (or user group’s) interest in or dependency on specific stations. The concept contributes to a better understanding of human mobility (e.g., travel purposes) and facilitates downstream applications, such as individual mobility prediction and location recommendation. However, intrinsic unsupervised learning characteristics and untrustworthy observation data make it challenging to estimate the real user-station attention. We introduce the user-station attention inference problem using station visit counts data in public transport and develop a matrix decomposition method capturing simultaneously user similarity and station-station relationships using knowledge graphs. Specifically, it captures the user similarity information from the user-station visit counts matrix. It extracts the stations’ latent representation and hidden relations (activities) between stations to construct the mobility knowledge graph (MKG) from smart card data. We develop a neural network (NN)-based nonlinear decomposition approach to extract the MKG relations capturing the latent spatiotemporal travel dependencies. The case study uses both synthetic and real-world data to validate the proposed approach by comparing it with benchmark models. The results illustrate the significant value of the knowledge graph in contributing to the user-station attention inference. The model with MKG improves the estimation accuracy by 35% in MAE and 16% in RMSE. Also, the model is not sensitive to sparse data provided only positive observations are used.

The increasing availability of travel trajectory data allows for a better understanding of travel behavior. In the individual mobility analysis, the problem of next trip prediction assumes a central role and is beneficial for applications such as personalized services and mobility management. This paper addresses the next trip prediction problem with arbitrary prediction times (the time when the prediction is made). This problem has not been studied adequately in the literature and it is important for applications driven by system events, such as proactive travel recommendations under disruptions or crowding in transport systems. It predicts an individual’s next trips given their historical trip sequences and the prediction time. We formulate the next trip prediction problem as on-board and off-board predictions depending on an individual’s travel status (i.e. on-board/off-board). Using historical/real-time travel trajectories, a DeepTrip model is proposed based on a trip sequence-to-sequence deep learning structure coupled with an attention mechanism. A novel overlapped embedding method is proposed to represent continuous travel attributes capturing simultaneously the categorical and numerical feature information. We also develop a random-sampling training algorithm to learn the impact of the prediction time. The model is validated using trip data in urban rails. The results show that DeepTrip outperforms statistical-based models by more than 10% in terms of accuracy and other deep learning models by 2%-3%. The impact analysis shows that different representations are appropriate for the two prediction cases (on-board/off-board), and the prediction performance does not monotonically improve as the prediction time approaches the next trip.