The operation of trains is substantially jeopardized by the presence of railway obstacles. Recent research on vision-based railway obstacle detectors has emphasized using sizeable neural network models to pursue a high accuracy while neglecting power consumption and training costs. To accelerate the model updating, training, and deployment cycle applied in the railway safety guarantee system, this article proposes a lightweight three-stage detection framework to identify obstacles in the single railway image. It is mainly comprised of the coarse region proposal (CRP) module, the lightweight railway obstacle detection network (RODNet), and postprocessing stage. In the CRP module, we set the binarized normed gradient (BING) detector to work on two different scale images for a more effective generation of proposal regions. Then, the intersection-over-minimum (IoM) metric is presented to filter and merge generated subregions. In the RODNet, a more lightweight and effective detection model is designed based on substantial improvements of YOLOv4-tiny, to simultaneously predict bounding boxes on the whole image and in each subregion. The RODNet’s backbone is rebuilt with one-shot aggregation and attention mechanism to reduce parameters and to maintain feature extraction abilities. Moreover, some residual blocks are also integrated into RODNet’s neck and head to improve the detection performance. Finally, a series of operations is devised to fuse, remove, and suppress predicted boxes on the boundary of each candidate subregion to reduce false-positive results at the postprocessing stage. Experimental results in the railway obstacle dataset show that RODNet reduces 48% parameters and 29% floating-point operations per second (FLOPs) with at least a 68.2% improvement of inference speed in the CPU compared with YOLOv4-tiny. Meanwhile, the proposed framework provides 80.3 mean average precision (mAP) at a speed of 65.9 ms/pic in the low-power graphics processing unit (GPU) and achieves a competitive or even better performance compared with the state-of-the-art large detection models.

Urban rail transit (URT) is often implemented in order to reduce car ownership and use. However, it is still unclear how the impacts of the built environment on car dependence differ in transit-oriented development (TOD) areas and non-TOD areas. Based on 2012 Changchun household travel survey data, this study aims to explore whether the relationships between the built environment and household vehicle kilometers traveled (VKT) vary between station and non-station areas by employing a multi-group structural equation model (SEM). The results suggest that the effects of the built environment on household VKT vary between station and non-station areas. Additionally, the results confirm the remarkable role that the built environment plays in influencing car ownership and use by revealing the direct, indiract and total effects. These findings provide a better understanding of the inherent mechanism by which URT moderates the impacts of the built environment on car ownership and use. It suggests that, for urban planners and policy makers, promoting compact land use strategies can have a positive effect on reducing car use, and URT implementation can effectively strengthen the effects.

Although the impacts of built environment on car ownership and use have been extensively studied, limited evidence has been offered for the role of spatial effects in influencing the interaction between built environment and travel behavior. Ignoring the spatial effects may lead to misunderstanding the role of the built environment and providing inconsistent transportation policies. In response to this, we try to employ a two-step modeling approach to investigate the impacts of built environment on car ownership and use by combining multilevel Bayesian model and conditional autocorrelation (CAR) model to control for spatial autocorrelation. In the two-step model, the predicting car ownership status in the first-step model is used as a mediating variable in the second-step car use model. Taking Changchun as a case study, this paper identifies the presence of spatial effects in influencing the effects of built environment on car ownership and use. Meanwhile, the direct and cascading effects of built environment on car ownership and use are revealed. The results show that the spatial autocorrelation exists in influencing the interaction between built environment and car dependency. The results suggest that it is necessary for urban planners to pay attention to the spatial effects and make targeted policy according to local land use characteristics.

Public transport services are often improved to decrease auto-related mobility. However, it remains unclear how the built environment influences auto ownership and commuting mode choice near public transit stations. Using 3318 individuals across China, this study contributes to investigating the relationship between the built environment, auto ownership and commuting mode choice simultaneously in neighborhoods near public transit stations by employing an integration of a structural equation model (SEM) and discrete choice model (DCM). The results indicate that the built environment exhibits significant impacts on auto-related mobility near public transit stations and auto ownership plays an intermediary role between the exogenous variables and commuting mode choice, thus it can moderate the influences of built environment attributes on commuting mode choice. The results suggest that increasing green coverage rates and promoting diverse land use development strategies in transit-oriented development (TOD) areas may make contributions to auto-related mobility reduction and low-carbon travel initiatives.