Accurate short-term passenger demand prediction contributes to the coordination of traffic supply and demand. This paper proposes an end-to-end multi-task learning temporal convolutional neural network (MTL-TCNN) predict in a multi-zone level. Along with feature selector named spatiotemporal dynamic time warping (ST-DTW) algorithm, this proposed MTL-TCNN is quite qualified for problem consideration correlations. Then, based on car-calling data from Didi Chuxing, Chengdu, China, taxi New York...

Despite the advancements in technologies of autonomous driving, it is still challenging to study safety a self-driving vehicle. Trajectory prediction one core function an This proposes Attention-based Interaction-aware Prediction (AI-TP) for traffic agents around With encoder-decoder architecture, AI-TP model uses Graph Attention Networks (GAT) describe interactions and Convolutional Gated Recurrent Units (ConvGRU) carry out predictions. Based on attention mechanism, constructs graphs from...

For autonomous vehicles driving on roads, future trajectories of surrounding traffic agents (e.g., vehicles, bicycles, pedestrians) are essential information. The prediction is challenging as the motion constantly affected by spatial-temporal interactions from and road infrastructure. To take those into account, this study proposes a Graph Attention Transformer (Gatformer) in which scene represented sparse graph. maintain spatial temporal information scene, Convolutional Neural Networks...

Birds-Eye-View (BEV) perception can naturally represent natural scenes, which is conducive to multimodal data processing and fusion. BEV contain rich semantics integrate the information of driving play an important role in researches related autonomous driving. However, constructed by single vehicle encounter certain issues, such as low accuracy insufficient range, thus cannot be well applied scenario understanding situation prediction. To address challenges, this article proposes a novel...

Abstract Knowledge of travel times serves an important role in traffic control and management. As increasingly popular data source, vehicle trajectories can provide large‐scale time information. However, real‐world information extracted from sparse or low‐resolution trajectory often contains missing that need to be imputed for further analysis. Thus, this study proposes a imputation generative adversarial network (TTI‐GAN) imputation. Considering the network‐wide spatiotemporal correlations,...

The structural efficiency of stiffened panels can be significantly improved by utilizing curvilinear stiffeners because their outstanding design flexibility. However, the explosion variables poses a stiff challenge to layouts such structures. In this study, novel layout optimization method is proposed for curvilinearly based on deep learning-based models, which enables intelligent design. Unlike traditional methods, image-based layout, characteristic stiffener paths, employed as variable,...

Though automated vehicles (AVs) are believed to play a crucial role in future transport, human driving will share the road with for relatively long period. So, we need enable run along drivers especially when they may have conflicts right of way. One key problem is how appropriately model behaviors and quickly simulate their actions training/testing vehicles. Many existing models were originally built traffic flow studies not be suitable studies. In this paper, propose set new principles...

Intelligent Transportation Systems (ITS) research and applications benefit from accurate short-term traffic state forecasting. To improve the forecasting accuracy, this paper proposes a deep learning based multitask Gated Recurrent Units (MTL-GRU) with residual mappings. enhance performance of MTL-GRU, feature engineering is introduced to select most informative features for Then, on real-world datasets, numerical results show that MTL-GRU can well estimate flow speed simultaneously,...

Due to the time-varying characteristics and interacted nature of multiple faults in high-speed train (HST), fault modeling, isolation, severity estimation cannot be described accurately using a single model, which may result poor performance conventional diagnosis methods. This article introduces idea models second-level adaptation techniques diagnose HST traction motor. First, reduced model description for is given. Then, isolation framework developed simplify parameters space segmentation....

ABSTRACT This article addresses the design of an observer‐based controller for networked Takagi‐Sugeno (T‐S) fuzzy systems with a two‐terminal adaptive event‐triggered mechanism under imperfect premise matching. First, novel scheme is proposed to optimize use limited communication resources. A observer variable different from that plant designed using non‐parallel distribution compensation (non‐PDC) strategy. Additionally, non‐PDC independent premises introduced enhance structural...

This study proposes a tensor-based K-Nearest Neighbors (K-NN) method, in which traffic patterns involve multi-dimensional temporal information and bi-directional spatial information. Such multi-temporal can not only capture the instantaneous fluctuation of short-term but keep general trend long-term traffic. In numerical experiments, with taxis' GPS data from an urban road network, speed are organized into one- (2 min), two- (4 min) three- (2, 4 10 dimensions. Meanwhile, about six upstream...

Coverage-guided Greybox Fuzzing (CGF) is one of the most successful and widely-used techniques for bug hunting. Two major approaches are adopted to optimize CGF: (i) reduce search space inputs by inferring relationships between input bytes path constraints; (ii) formulate fuzzing processes (e.g., transitions) build up probability distributions power schedules, i.e., number generated per seed. However, former subjective inference results which may include extra a constraint, thereby limiting...

In this paper, we systematically study how to use edge computing monitor the movements of multiple connected and automated vehicles (CAV) warn potential accidents (e.g., lane departures, collisions). Compared conventional approaches that only sensing data individual vehicles, cooperative vehicle infrastructure systems directly collect movement via vehicle-to-everything (V2X) communications thus easily calculate risk every synthetically. We propose a fast algorithm corresponding structure...