Efficiently updating real-time road maps is essential for autonomous driving and significantly influences human driving decisions. Existing high-definition (HD) maps are often based on static data and neglect real-time road information, particularly temporary traffic control details related to construction. In this study, we propose a method to identify and update lane-level information for roads in restricted areas using only a digital video recorder and a low-cost global navigation satellite system (GNSS). This research proposes a Road Temporary Traffic Control Network (RTTCNet) to detect control devices and lane markings and to optimize a 3D reconstruction method that can accurately position control devices on HD maps. The method identifies areas with restricted road access and analyzes the spatial relationship between control devices and lane lines to update the HD map. Our method effectively addresses the high costs and inefficiencies associated with conventional map-updating methods. Experimental tests on various road types in Shanghai demonstrated 94.85% accuracy in identifying restricted areas for road access from single images, with a mean positioning error of 3.86 meters for temporary traffic control devices on HD maps. This method enables real-time lane-level updates for road information on HD maps and more effective decision-making support for autonomous driving.

Sampling timestamps enhance the reliability of spatiotemporal trajectory density estimation. Existing kernel density estimation (KDE) methods leverage sampling time to modify kernel shapes or weight specifications. However, controlling two concurrent errors remains challenging: density over/underestimation due to temporal autocorrelation, and nonzero density assignments to unreachable locations. This study proposes a time-aware KDE (t-KDE) method for irregular sampling intervals, adopting a divide-and-conquer strategy that synergistically integrates time geography and temporal autocorrelation. By establishing two temporal mappings—from time to kernel functions and to weight coefficients—t-KDE retrospectively mitigates the above errors through KDE recalibration. The approach aims to maximize temporal information utilization to reduce estimation uncertainty, providing a theoretical basis for unbiased density modeling. Empirical analysis demonstrates that t-KDE outperforms state-of-the-art methods in accuracy and reliability.

Vehicle trajectory prediction (VTP) is a prominent research topic in autonomous vehicles (AVs). However, existing studies concentrate on improving prediction accuracy, overlooking how VTP accuracy influences the effectiveness of downstream tasks such as collision avoidance. In this paper, we investigate the relationship between VTP accuracy and key performance indicators (KPIs) of cooperative collision avoidance systems (CCAS) for connected AVs operating at blind intersections. To this end, we propose a generalized error model (GEM) that enables the generation of customizable VTP errors. GEM is integrated into an error injection framework to produce diverse, error-controlled trajectories in real time. Our experiments quantitatively characterize the relationship among VTP accuracy, warning thresholds and strategies, and CCAS KPIs. The results reveal a common pattern: as VTP error increases, CCAS recall improves while precision declines, demonstrating a negative correlation that is difficult to balance. This underscores the need for effective trade-offs between recall and precision in CCAS design. With a general collision warning strategy, CCAS can preemptively report over 90% of collisions, substantially enhancing traffic safety, though at the expense of a reduced precision. To mitigate this, we devise a density-based warning strategy (DCWS) that improves precision across the entire VTP accuracy space while maintaining recall at acceptable levels. Notably, DCWS gradually decouples CCAS KPIs from VTP accuracy performance, highlighting the crucial role of warning strategies. The quantitative results presented in this paper can help practitioners assess the impact of their VTP models on CCAS and select appropriate warning thresholds.

High-definition (HD) maps are becoming increasingly important for autonomous vehicles and advanced driver assistance systems (ADAS), as they provide detailed and accurate information about the road environment. Vision localization using HD maps can improve accuracy, but maps can become outdated. Image perception can be useful for HD map change detection. However, achieving robust and high-precision alignment between images and HD maps is challenging in varying environmental conditions. In addition, evaluating alignment in the absence of ground truth data is inconvenient. This article proposes a trajectory interpolation-based method for reconstructing lane markings from images to detect road changes when compared with the HD grid map. Images and HD maps are registered using a distance transform and cross-entropy-based optimization. The article also proposes a metric based on intersection over union for evaluating alignment accuracy. Experiments were conducted using a third-party-collected campus dataset and the Argoverse2 open-source dataset to demonstrate the effectiveness of the proposed methods in detecting road changes and achieving high-precision image and HD map fusion.

Although the effects of spatial abilities on behaviour have recently been a hot topic, wayfinding has received little attentions. Here, the reasons for route selection differences during wayfinding were studied from the perspective of individual competence. The mental rotation (A-MR), abstract reasoning (A-AR), visual short-term memory (A-VSTM), spatial perception (A-SP) and spatial orientation (A-SO) abilities of 52 college students were assessed, and their route selection behaviours were tested in homogeneous and heterogeneous three-dimensional virtual scenes. Finally, the relationships between five abilities and the lengths (R-L) and angles (R-A) of the selected routes were analysed and the R meants route here. The path complexity (weighted combination of R-L and R-A) was also fitted using a partial least-squares model with the five abilities. The A-MR and A-AR had obvious relationships with R-L and R-A, whereas A-SP showed little correlation. Moreover, the A-VSTM and A-SO had a significantly higher impact on route selection in the homogeneous spatial environment than in the heterogeneous one. Although the models passed the significance test, the fitting effect was not satisfactory. The routes were affected by many factors, and these five abilities could not fully explain the route selection behaviour of the individuals.