Foreign object intrusion into high-speed railway (HSR) catenary systems poses severe operational hazards, making effective detection crucial for safety. Precise detection of these small intrusive objects is essential. However, the lack of datasets and research on foreign object intrusion in HSR scenario brings two major challenges: limited data and low accuracy for detecting small intrusive objects. To address these challenges, this paper introduces a novel generative method for detecting foreign object intrusion. To address data limitations, we use low-rank adaptation to fine-tune a diffusion model, developing a generation-extraction-integration framework that generates true-to-reality HSR images of small intrusive target objects. Furthermore, to enhance the detection of small objects in HSR scenario, we propose a new detection model called SA-YOLO. Based on the YOLOv9 architecture, this model optimizes the backbone network using the star operation, an element-wise multiplication method, and introduces the A-DyS module to improve upsampling through dynamic sampling and attention mechanism. Extensive experiments demonstrate that in the HSR scenario our method outperforms existing state-of-the-art approaches in terms of both generation quality and detection performance, while also showing high robustness.

This article proposes a new multivehicle cooperative localization approach that combines time of arrival (TOA) with a heuristic simulated annealing extended Kalman filter (SA-EKF) to enhance positioning accuracy and robustness in urban canyons. The method incorporates a path loss model to account for the complex communication environment between vehicles, using TOA measurements for distributed EKF estimation. The integration of a simulated annealing strategy within the method is instrumental in circumventing local minima, thereby facilitating global optimisation. Furthermore, an adaptive weighted filtering correction mechanism is employed to enhance estimation accuracy and system stability. Experimental results conducted on the SUMO simulation platform and in real-world scenarios demonstrate that the proposed method offers certain advantages over existing approaches in complex, noisy environments.

We present a novel solution for mesh-based deformation simulation from a spectral perspective. Unlike existing approaches that demand separate training for each garment or body type and often struggle to produce rich folds and lifelike dynamics, our method achieves the quality of physics-based simulations while maintaining superior efficiency within a unified model. The key to achieve this lies in the development of a spectrum-enhanced deformation network, a result of in-depth theoretical analysis bridging neural networks and garment deformations. This enhancement compels the network to focus on learning spectral information predominantly within the frequency band associated with intricate deformations. Furthermore, building upon standard blend skinning techniques, we introduce target-aware temporal skinning weights. The weights describe how the underlying human skeleton dynamically affects the mesh vertices according to the garment and body shape, as well as the motion state. We validate our method on various garments, bodies, and motions through extensive ablation studies. Finally, we conduct comparisons to confirm its superiority in generalization, deformation quality, and performance over several state-of-the-art methods.

Understanding decision-making in autonomous driving models is essential for real-world applications. Attribution explanation is a primary research direction for interpreting neural network decisions. However, in the context of autonomous driving, numerical attributions fail to interpret the complex semantic information and often result in explanations that are difficult to understand. This paper introduces a novel semantic attribution approach that both identifies where important features appear and provides intuitive information about what they represent. To establish the semantic correspondences for attributions, we propose an interpreting framework that integrates unsupervised differentiable semantic representations with the attribution computational model. To further enhance the accuracy of the attribution computation while ensuring strong semantic correspondence, we design a Semantic-Informed Aumann-Shapley (SIAS) method, which defines a novel integration path solution using constraints from semantic scores and discrete gradients. Extensive experiments confirm that our method outperforms state-of-the-art explanation techniques both qualitatively and quantitatively in autonomous driving scenarios.

Given the critical need for more reliable autonomous driving systems, explainability has become a key focus within the research community. In autonomous driving models, even minor perception differences can significantly influence the decision-making process, and this impact often diverges markedly from human cognition. However, understanding the specific reasons why a model decides to stop or keep forward remains a significant challenge. This paper presents an attribution-guided visualization method aimed at exploring the triggers behind decision shifts, providing clear insights into the underlying “why” and “why not” of such decisions. We propose the cumulative layer fusion attribution method that identifies the parameters most critical to decision-making. These attributions are then used to inform the visualization optimization by applying attribution-guided weights to crucial generation parameters, ensuring that decision changes are driven only by modifications to critical information. Furthermore, we develop an indirect regularization method that increases visualization quality without necessitating additional hyperparameters. Experiments on large datasets demonstrate that our method produces insightful visualization explanations and outperforms state-of-the-art methods in both qualitative and quantitative evaluations.

Deep neural networks (DNNs) have advanced autonomous driving, but their lack of transparency remains a major obstacle to real-world application. Attribution methods, which aim to explain DNN decisions, offer a potential solution. However, existing methods, primarily designed for image classification models, often suffer from performance degradation and require specialized algorithmic adjustments when applied to the diverse models in autonomous driving. To address this challenge, we introduce a universally applicable representation of traffic scenes, forming the basis for our unified attribution method. Specifically, we leverage the first-order Taylor expansion at a specific hidden layer, i.e., the product of gradients and feature maps, to represent abstract traffic scene information. This representation guides both the optimization of attribution path generation and the attribution computation, enabling consistent and effective attributions for both lane-change prediction and vision-based control models. Experiments on two distinct autonomous driving models demonstrate that our approach outperforms state-of-the-art methods in explanation accuracy and robustness, advancing the interpretability of DNN-based autonomous driving models.

Recent research has demonstrated that Vision Transformers (ViTs) are capable of comparable or even better performance than convolutional neural network (CNN) baselines. The differences in their structural designs are obvious, but our understanding of the differences in their feature representations remains limited. In this work, we propose several techniques to achieve high-quality visualization of representations in ViTs. Both qualitative and quantitative experiments show that our technical improvements can observably improve ViT visualization quality compared to previous studies. Furthermore, we conduct visualizations to explore the disparities between ViTs and CNNs pre-trained on ImageNet1K, revealing three intriguing properties of ViTs: a) ViT feature propagation retains image detail information with minimal loss, whereas CNNs discard most image details for class discrimination. b) Different from CNNs, object-related features do not show in ViT higher layers, suggesting that class-discriminative features may not be required for ViT classification. c) Our visualization-assisted texture-bias experiment reveals that both ViTs and CNNs exhibit texture bias, of which ViTs seem to be more biased towards local textures.

Parallelization is widely employed to improve the exploration ability of controllers. However, it is rare to provide a lightweight scheme for reducing homogeneous policies with theoretical guarantees. This article is concerned with a novel parallel scheme for solving optimal control problems. In brief, we design a novel global indicator that inherits the theoretical guarantees of a class of iterative reinforcement learning algorithms. By generating a tentative function, the global indicator can guide and communicate with parallel controllers to accelerate the learning process. Using two typical exploration policies, the novel parallel scheme can rapidly compress the neighborhood of the optimal cost function. Besides, two parallel algorithms based on value iteration and Q-learning are established to improve the data efficiency through different extensions. Finally, two benchmark problems are presented to demonstrate the learning effectiveness of the novel parallel scheme.