Computer vision techniques have significantly advanced autonomous driving but still face specific challenges in complex environments, such as accurately detecting small, occluded, or low-contrast objects amidst dynamic backgrounds and varying lighting conditions. This paper introduces FOS-YOLO, a novel model designed to enhance detection performance in these scenarios. FOS-YOLO integrates the Focal Nets mechanism to improve feature representation, effectively addressing target scale variations and low contrast through multi-scale context aggregation and multi-level modulation. Additionally, the Zoomed Spatial Convolutional Block Attention Module (ZS_CBAM), is introduced to improve the detection of small targets, occlusions, and illumination changes by effectively fusing channel and spatial attention with a focus on scaled spatial features. The model also includes two lightweight modules, the Adaptive Depthwise Enhanced Module (ADEM) and the Lite Enhanced Reduction Module (LERM), which reduce parameters and computational load, accelerating convergence and improving accuracy. Experimental results, compared with state-of-the-art methods, show that FOS-YOLO achieves a mAP of 90.4% on the KITTI dataset and 64.3% on the RTTS dataset, with a 20.07% reduction in parameters, significantly enhancing real-time detection accuracy and efficiency.

The vibration degree of a steering wheel has important reference significance for drivers to evaluate the ride comfort of the whole vehicle. To solve the jitter problem of the steering wheel of a commercial vehicle at idle speed, this work proposes a multinode joint vibration control strategy (MDVC) based on the associated vibration path of the steering wheel. Based on the analysis of the associated vibration transfer paths of the steering wheel, the whole vehicle was divided into a system comprising several nodes. For the decomposed node system, taking the vibration transmission path associated with the target as the research direction, the vibration reduction design of each node system is analyzed step by step. After exploring the possible causes of abnormal vibration of the steering wheel through experimental tests, the abnormal node structure interval was determined. By further extracting the structural model of the steering system from the vehicle, the hammering method was applied to test its modal and related frequency. Furthermore, an improved structure of steering support was also designed, and its fitting degree and modal characteristics were analyzed and compared to the original scheme. The following test results show that the structure improvement greatly reduces the vibration level of the steering wheel, meets the ideal design requirements of the steering wheel vibration reduction, and provides the possibility of weighing the correlation between these hierarchical node systems in whole vehicle.

Reliable perception under adverse conditions remains a fundamental challenge for intelligent transportation systems. To address the limitations of vision-based detection in complex traffic scenarios, this paper proposes DSG-YOLO, an efficient traffic-scene adaptive attention network, together with a robust decision-level visual–radar fusion strategy. The design explicitly targets four real-world pain points: crowding-induced localization ambiguity, small and occluded objects, multi-scale feature mismatch, and inconsistent cross-modal cues. Specifically, a lightweight Grouped Spatial-Aware Attention Module (GSAM) is introduced to mitigate localization ambiguity in cluttered environments. A Synergistic Channel–Spatial Attention (SCSA) module combined with Omni-dimensional Dynamic Convolution (ODConv) enhances detection of small and occluded objects via adaptive multi-scale fusion. To achieve consistent multi-modal association, a decision-level fusion framework based on Intersection-over-Union (IoU) and Global Nearest Neighbor (GNN) matching is proposed. Experiments on KITTI, RTTS, and real-vehicle platform demonstrate that DSG-YOLO achieves 90.1% mAP on KITTI and 55.3% on RTTS. On real-vehicle data, the fusion improves precision by 3.11%, reduces false positives by 1.61%, and enhances robustness of real-time physical measurements, demonstrating its potential for intelligent transportation.

Investigating the rubbing failure characteristics in turbine rotors holds significant implications for ensuring the operational integrity and reliability of aero-engine systems. A mixed-nodes boundary feature method (MNBFM) was proposed to investigate the dynamic characteristics of rotors experiencing rubbing failure. In this approach, the key components of the turbine pump are modelled as flexible bodies, with their surface profile nodes and connection nodes systematically organized and integrated. Then, the dimension reduction of the components is realized through the Craig-Bampton method, and the dynamic characteristics of the turbine rotor under various dynamic rub-impact parameters and loading conditions are investigated. The results reveal that: (1) The optimized model significantly reduces file size and computation time while maintaining high response accuracy. (2) the trajectory of the axis after rubbing exhibits a noticeable clipping phenomenon. (3) Higher rotational speed intensified the rubbing frequency and increase the amplitude of individual rubbing events. In addition, the rubbing force shows the intermittent characteristic of jump attenuation.

The vibration level of a cab affects the passenger’s ride comfort and safety significantly. It is of great importance to control the vibration level of cabs under various driving conditions. The associated vibration transfer paths of cabs are studied by using a hierarchical analysis method of a parameter index. The multiobjective design analysis is carried out by using the multiparameter joint optimisation design. Further, the optimal control of the cab vibration level is obtained from a full-condition simulation environment. Additionally, a multibody vehicle model is established. The simulation analysis under multiple working conditions is conducted. The optimal parameter distribution of the cab mounting structure was established by analysing the influence of the design parameters and experimental verification. This greatly improves the comfort of the cab.

Accurate alignment between subjective perception and objective measurement is critical for evaluating ride comfort in commercial vehicles. Traditional assessment methods are often time-consuming, subjective, and lack a quantitative connection between human perception and engineering parameters. To address these limitations, a regression-based predictive model is proposed, utilizing a three-tier hierarchical framework for commercial vehicle comfort evaluation. This framework decomposes overall vibration comfort into three levels: first-level indicators (overall comfort), second-level indicators (primary ride comfort, secondary ride comfort, and impact harshness), and ten third-level quantifiable sub-indicators, such as vertical motion, lateral shake, and residual vibration. Pearson correlation analysis was used to identify critical measurement points, with the driver's seat rail showing the strongest and most consistent correlations with subjective scores across all indicators. A systematic mapping was established to link each subjective indicator to specific vibration parameters at the seat rail, including single-axis vibrations, combined multi-axis responses, peak values, and decay time. Ridge regression addressed multicollinearity among objective parameters, resulting in high predictive accuracy, strong goodness-of-fit, and low prediction errors across all indicators. Furthermore, intermediate models tailored to road-specific conditions, distinguishing between random and impulsive inputs, were developed to enable comprehensive ride comfort assessment within a unified predictive framework. Validation results indicate that the proposed model effectively translates subjective feedback into quantifiable engineering parameters for vehicle Noise, Vibration, and Harshness (NVH) performance evaluation, substantially reducing reliance on labor-intensive subjective testing and minimizing evaluator bias. This approach offers commercial vehicle manufacturers a practical and cost-effective tool for suspension design optimization, NVH refinement, and reduced vehicle development cycle time.