Search Results (638)

Mountain-greenway planning requires approaches that account for heterogeneous user needs, trail safety, experiential quality, and limited maintenance capacity. Taking the Shenzi Village–Longshan Village section of Wugong Mountain as a case, this study develops a perceived-value-informed, segment-level, scenario-based framework for assessing mountain greenway suitability. COROS PACE 3 FIT trajectories, field observations, and researcher-based environmental interpretation were used to construct eight indicators for ten segments. Entropy weights characterize within-sample variation, while the Potential Suitability Value (PSV) integrates environmental conditions with a priori utility rules for low-, moderate-, and high-activity scenarios. Ranking sensitivity was examined through utility-function sampling, fixed-endpoint comparisons, equal-weight comparisons, leave-one-segment-out reweighting, weight perturbation, and indicator-deletion analysis. Shade, exposure, movement difficulty, and cumulative ascent exhibited relatively high differentiation within the present sample. S10 ranked among the top three under all three scenarios in both the baseline entropy-weight and leave-one-segment-out schemes, but fell to fourth under the high-activity scenario with equal weights. The low-activity rankings of S5 and S8 were jointly affected by the span and internal spacing of the utility functions, whereas the high-activity advantage of S6 depended structurally on the treatment of risk indicators. By identifying segments with inadequate environmental support, elevated safety pressure, or parameter-sensitive rankings, the framework can inform targeted field verification, maintenance prioritization, and risk communication. It is a scenario-based suitability and planning-support tool, not an assessment of sustainability outcomes. Ecological disturbance, environmental carrying capacity, community benefits, social equity, long-term maintenance costs, resource efficiency, and actual user perceptions were not measured and require separate indicators and external validation. Full article

Comprehensive evaluation of open-pit blasting is challenging because performance indicators originate from different sensing sources and are predicted with different levels of model performance. This study proposes a multi-source sensing and interpretable ensemble learning framework for model-performance-weighted blast-performance evaluation. Data from 70 production blasts were organized into 12 input variables and 10 indicators spanning fragmentation quality, muckpile morphology, safety and adverse effects, and operational efficiency. Random forest, XGBoost, LightGBM, and CatBoost models were developed separately for each indicator and compared using five-fold cross-validation, while SHAP was applied to interpret the selected models. The indicator-specific models achieved cross-validated R² values of 0.7812–0.9012; these are selection-conditioned cross-validated estimates obtained under the same five-fold partition that was used to select the model for each indicator, and they are not unbiased estimates of generalization performance. Powder factor and uniaxial compressive strength were influential for fragmentation, whereas maximum charge per delay strongly affected peak particle velocity and muckpile displacement. Cross-validated R² was then incorporated into AHP–entropy weighting as a relative model-performance coefficient that is not a confidence level, a probability of correctness, or an uncertainty estimate. In field application, engineering adjustment based on model feedback increased the comprehensive score from 65.4 to 87.5, improved the grade from III to II, and reduced D80 and PPV by 25.4% and 27.8%, respectively. The models are specific to the monitored mine rather than transferable, and the field implementation is an application case rather than an independent external validation. The framework provides an interpretable and model-performance-aware basis for site-specific blast-scheme comparison and iterative improvement. Full article

Dooring accidents remain a major urban traffic safety hazard, often caused by insufficient visual attention to the lateral environment before door opening. Despite growing research on driver hazard perception, how the approach speed of lateral vulnerable road users modulates drivers’ visual attention allocation and risk perception–decision-making in stationary pre-door-opening scenarios remains poorly understood. This study investigates the visual and decision-making responses of stationary drivers under lateral threats of varying speeds. A static real-vehicle experimental platform was established to reproduce a natural roadside parking scenario, in which electric bicycles approached from the left rear at five constant speeds ranging from 5 to 25 km/h. Eye movement data from 20 drivers were collected and analyzed using a Markov transition model, Shannon entropy, and one-way ANOVA to extract Area of Interest (AOI) fixation, gaze sequence entropy, and decision time. The results show that drivers’ visual strategies exhibited a strong non-linear modulation by approach speed, shifting from direct window observation at low speeds, to structured mirror-based checking at intermediate speeds, and to distributed scanning under the highest speed; gaze sequence entropy decreased to a minimum at 20 km/h and rebounded at 25 km/h. Although mean decision time decreased with increasing speed, spatial–temporal analysis revealed that at higher speeds drivers made the “Dangerous” judgment only after the electric bicycle had already passed them, showing that a shorter decision time does not reflect a more effective response. These findings offer a methodological reference for future research on driver attention and risk perception–decision-making in pre-door-opening scenarios. Full article

Sequential evidence selection can reveal which parts of an existing sensor record reduce model uncertainty, while providing a computational test bed for future adaptive sensing. This work introduces the Embodied Evidence Acquisition and Reasoning Loop (EARL), a typed architecture in which five role-specialized critics score evidence requests and a deterministic safety governor retains exclusive execution authority. Observed, derived, and simulated evidence remain provenance-distinct in a replayable, hash-linked ledger. The primary experiment sequentially disclosed precomputed feature bundles from pre-recorded gait data; it did not acquire new measurements. EARL was evaluated on 64 unique subjects from the PhysioNet Gait in Neurodegenerative Disease Database using repeated subject-level cross-validation, 11 predeclared conditions, and 3520 replay-verified runs. The primary endpoint was area under cumulative posterior-entropy reduction. EARL achieved 7.689 (95% CI 7.535–7.840), exceeding fixed-order and random selection by 0.459 and 0.763, respectively; the EARL-minus-EIG difference was –0.321. EARL had descriptively higher macro accuracy (55.8% versus 49.5%) and a lower Brier score (0.781 versus 0.813) than pure expected-information-gain selection. This measures an entropy-efficiency trade-off under additional selection criteria, not universal superiority. All 10 original software safety challenges produced their expected outcomes. Separately identified post hoc analyses examine probe use, a sensitivity analysis excluding the record-quality probe, critic influence, illustrative resource costs, early stopping, and safety-constrained simulated execution. The small retrospective cohort and illustrative simulations establish software behavior, not clinical diagnostic performance, treatment benefit, physical-robot safety, or patient efficacy. Full article

Threshold-based flight data monitoring (FDM) identifies parameter exceedances but provides limited visibility into gradual operational deviation within nominal operating bounds. This study presents a three-layer framework for continuous operational deviation quantification from quick access recorder (QAR) data. A standard operating procedure (SOP)-constrained Gaussian mixture model–hidden Markov model (GMM-HMM) infers 27 procedural states from flight parameters, aligning latent state representation with operational procedure. From these state sequences, maximum entropy inverse reinforcement learning (IRL) recovers an expert value function that estimates execution quality at each time step. Split conformal prediction calibrates detection thresholds with distribution-free coverage guarantees, while counterfactual analysis traces identified deviations to their procedural origin. Learned from real flight data, the recovered reward function assigns 1.6 times greater weight to lateral and speed stability than to vertical stability. Split conformal calibration yields an empirical false alarm rate of 4.2% under a nominal 5% significance level. Counterfactual analysis identifies vertical speed management as the dominant source of procedural deviation. These results support the proposed framework as a candidate continuous safety performance indicator that surfaces gradual operational deviation that threshold-based exceedance monitoring would not flag. Full article

Vehicular networks require privacy-preserving and secure authentication mechanisms to protect safety-critical communications against evolving cyber threats. Conventional authentication schemes relying on elliptic-curve cryptography (ECC) and RSA are vulnerable to quantum attacks, making them unsuitable for next-generation intelligent transportation systems. This paper proposes a unified post-quantum pseudonymous authentication and key establishment (UPQ-PAKE) scheme for secure vehicular networks. The proposed framework integrates dynamic pseudonymous identity construction, post-quantum digital signatures, and dual ephemeral key encapsulation into a single transcript-bound authenticated key exchange protocol. ML-DSA is employed for mutual authentication, while ML-KEM enables quantum-resistant session key establishment. Dynamic session-specific pseudonyms provide identity privacy and unlinkability. Furthermore, transcript binding, nonce freshness verification, and contributory dual-ephemeral session entropy strengthen the protocol against replay attacks, provide forward secrecy, and resist man-in-the-middle attacks under the quantum polynomial-time adversarial model. Formal security analysis demonstrates that the proposed scheme achieves authenticated key exchange security based on the IND-CCA security of ML-KEM and the EUF-CMA security of ML-DSA. Performance evaluation demonstrates that the proposed authentication technique maintains practical computational and communication overhead and provides post-quantum mutual authentication, dual-directional key establishment, and dynamic unlinkability, making it suitable for large-scale, real-time IoV deployments. Full article

Laser strapdown inertial measurement units (LIMUs) are critical components of autonomous navigation and attitude determination systems used in mission-critical platforms such as spacecraft. Their performance directly affects the reliability and safety of space missions. To address the challenges posed by high-dimensional indicator sets, limited data, and stringent interpretability requirements, this paper proposes an interpretability-aware hierarchical Bayesian belief rule base model, termed IH-BRB, for LIMU performance evaluation. First, a five-level hierarchical evaluation structure is developed based on the physical architecture and error-propagation relationships of a LIMU, thereby reducing rule base complexity and improving inference traceability. Second, an interlayer transmission mechanism based on the expected grade scores and entropy-derived reliability is introduced to preserve essential uncertainty information and dynamically adjust the attribute weights of upper-level BRB nodes. Third, semantic consistency constraints are imposed on the conditional probability tables and rule-consequent belief distributions to prevent semantic drift during data-driven optimization. The proposed model is validated using an engineering LIMU dataset with expert-defined target grades to assess its consistency with the established grading criterion. The results show that IH-BRB maintains a competitive evaluation performance while preserving physically meaningful rule semantics and layer-wise traceability. Full article

Musculoskeletal injuries constitute a critical industrial challenge, yet predictive ergonomic methodologies have remained largely stagnant. Conventional assessments are primarily limited to static, linear snapshots of posture, resulting in a failure to capture the inherently dynamic and chaotic nature of human movement over time. To address this limitation, the present study introduces an innovative approach grounded in nonlinear dynamical systems and robust statistics. The objective was to determine whether information-theoretic and dynamic metrics, specifically Shannon Entropy and the Largest Lyapunov Exponent (LyE), can accurately quantify the structural stability of the cervical spine during a fatigue-inducing isometric protocol. To mitigate the inherent noise of field-derived biomechanical data and prevent algorithmic singularities during phase-space reconstruction, zero-phase Butterworth filtering as well as uniform Gaussian noise injection (Gaussian dither) were systematically applied. The analysis revealed a clear, quantifiable topological shift: as localized muscular fatigue accumulates, the biomechanical system abruptly transitions from a state of stable, organized kinematic flow directly into chaotic instability. Furthermore, these interpretable nonlinear metrics demonstrate high suitability for integration into Explainable Artificial Intelligence (XAI) safety architectures. By transforming complex, stochastic biomechanical data into clear diagnostic biomarkers, this research establishes a foundational framework for proactive, real-time preventive safety measures, thereby advancing the paradigm of Syntropic Bioergonomics. Full article

Aviation electromagnetic (EM) interfaces present complex, time-varying information and require designs that minimize unnecessary cognitive demands. However, evidence on how visual encoding and spatial layout affect performance in dynamic EM tasks remains limited. This study compared two dimensions of visual encoding (radar-range and target-threat) and two spatial layouts using a 2 × 2 × 2 repeated-measures design with 32 participants. A Unity 3D simulation incorporated concurrent target recognition and EM situation monitoring tasks. Reaction time, accuracy, NASA Task Load Index (NASA-TLX), interface preferences, and eye-tracking metrics were collected. Target-threat encoding significantly affected target-recognition accuracy, with border-thickness threat encoding outperforming color-background encoding, whereas reaction time did not differ significantly. Radar-range encoding significantly influenced EM monitoring, with the grid encoding producing shorter reaction times than the color-overlay. Target-threat encoding and spatial layout significantly affected subjective workload. The grid-based radar sector combined with border-thickness threat encoding and the center layout was the most frequently preferred interface. Eye-tracking metrics and heatmaps further indicated that the center layout facilitated faster visual attention to the EM indicator, accompanied by fewer total fixations and lower gaze entropy after anomaly onset. These findings support perceptually separable target encoding and task-centered placement of high-priority EM information may inform future helmet-mounted and other wearable displays for dynamic, safety-critical tasks. Full article

The sustainability of transport planning in intermodal route selection depends on the examination of a set of criteria and alternatives. This paper investigates a dynamic decision-making process by studying the change in ranking when a new alternative is added, or an existing alternative is removed; or when a new criterion is added, or an existing one is removed. The presence of a change in ranking reveals the phenomenon of rank reversal (RR) in multi-criteria analysis. The Sequential Iterative Method for Urban Systems multi-criteria method (SIMUS) is presented as a tool to study the rank reversal problem. The methodology of the research consists of six steps. In the first step, the initial decision-making matrix is compiled. The second step includes the compilation of new decision-making matrices by adding or removing alternatives from the initial decision-making matrix. In the third step, the SIMUS method is applied. In the fourth step, the influence on rank reversal of the addition of alternatives that are very similar or identical to another existing in the system is analyzed. In the fifth step, the influence of a change in the number of criteria on the rank reversal is studied. The sixth step includes determination of the entropy and information of the system. A case for the sustainability of container carriage by railway and road routes between Sofia and Varna in the Bulgarian transport network is demonstrated. Three railway routes and three road routes served by different types of locomotives and heavy trucks have been studied. The routes were evaluated according to the following criteria: CO₂ emissions, operating costs, infrastructure fees, travel time, route length, safety and security factor of transport. The change in the ranking of alternatives was studied by comparison of rankings for cases from 3 to 12 alternatives. A hypothesis based on the assumption that RR in the SIMUS method is a logical consequence of linear transformations along different dimensional spaces, each time an alternative is added or deleted, has been introduced. An invariance coefficient and a rank reversal coefficient have been defined. The values of both coefficients have been classified into groups to explain sustainability in decision-making. Full article

An autonomous underwater vehicle (AUV) is an important operation platform in ocean exploration engineering, and the operation state of its actuator directly determines the navigation safety. However, the marine environment is complex and changeable, and the state signal characteristics of autonomous underwater vehicles are complex and vulnerable to noise interference. The performance of the traditional fault identification model degrades significantly in the noise environment, and the model ignores the risk difference of misjudgment, which may lead to serious misjudgment consequences. In order to solve the above problems, we propose a Risk-Sensitive Dual-Branch Learnable Filtering (RS-DBLF) framework for AUV fault diagnosis. The feature extraction backbone network adopts a dual-branch structure, which can adaptively capture the low-frequency slow degradation fault features and high-frequency impact fault features, and simultaneously complete the noise suppression and multi-scale fault frequency domain feature separation. Secondly, a fault asymmetric cost matrix is established for the cost-sensitive diagnosis scenario, and the expected misjudgment loss corresponding to various faults is calculated by using the posterior probability of the network output. The global minimum-risk criterion is used to replace the traditional cross entropy hard classification reasoning, so as to reduce the high-cost fault misjudgment probability from the decision level. The “Haizhe” open dataset is used to verify the effectiveness of the algorithm. The results show that the cost-sensitive dual-branch learning bandpass network has a stronger robustness in noise environments, and the minimum-risk decision based on Bayesian can effectively reduce the loss of misjudgment. Full article

Human capital represents the core asset upon which the social dimension of the transition towards the circular economy is based, as the strategies for transition depend on people and their skills, safety, development and inclusion that ultimately determine the success in meeting the human-capital-related Sustainable Development Goals. Human capital, however, represents the least measured dimension of corporate sustainability, since the existing practices of measuring human capital involve either its monetary evaluation or subjective weight allocation within the firm-specific sustainability scorecard or aggregation at the national level without the ability to create an objective firm-level benchmark of human capital performance. In this study, human capital represents a management object, and the question is about the extent to which the standardized public disclosure of data about employment, occupational health and safety, training and education, and diversity and equal opportunity according to the Global Reporting Initiative standards could be transformed into a composite index that would discriminate managerial performance of organizations in managing the human capital-related SDGs. For this purpose, the Managerial Performance Index, a composite indicator based on a data-driven approach to creating a benchmark, where indicators are normalized, weighted objectively by means of Shannon entropy for their discriminating capacity and aggregated into four sub-indexes and one overall score, is created. The framework is applied to the cross-sectional sample of the leading European airport operators—an industry characterized by intensive capital, safety and labor intensity but never benchmarked in terms of human capital before. Full article

Deer Horn Glue is rich in collagen- and tissue-derived proteins, but the link between its experimentally observed proteome, the theoretical peptide sequence space derived from that proteome, and bone-related molecular targets remains poorly defined. This study established a proteomics-guided multiscale computational workflow to prioritize putative receptor activator of nuclear factor-κB ligand (RANKL)-binding peptide candidates generated from proteins identified in one Deer Horn Glue sample. The Deer Horn Glue proteome was characterized by liquid chromatography–tandem mass spectrometry (LC-MS/MS). Experimentally identified proteins were subjected to in silico tryptic digestion and stepwise activity, safety, and physicochemical screening. All 49 retained candidates underwent global and interface-focused docking. Five prioritized peptides were examined by AlphaFold 3, three independently seeded 200 ns molecular dynamics simulations per complex, entropy-omitted molecular mechanics/generalized Born surface area (MM/GBSA) analysis, residue decomposition, and locally relaxed computational alanine substitution analysis. Proteomic analysis retained 707 target protein groups and generated 24,575 nonredundant theoretical peptide sequences. Stepwise screening retained 49 candidates. Global and interface-focused docking rankings showed modest agreement, and the expanded interface analysis identified additional candidates while retaining GASLQDWDFGK as the top-ranked sequence. Across three independent simulations, GASLQDWDFGK showed consistently low peptide root-mean-square deviation (RMSD), whereas HEFSVDMTCEGCSNAVTR formed the largest mean number of interfacial hydrogen bonds. Their entropy-omitted MM/GBSA estimates were generally the most favorable, although the order varied among simulations. Locally relaxed alanine substitutions highlighted reproducible energy-sensitive positions for experimental testing. By linking an experimentally observed Deer Horn Glue proteome to multiscale structural analysis, this workflow provides a traceable and reproducible strategy for prioritizing testable peptide–RANKL interaction hypotheses. The five candidates remain theoretical products of in silico digestion and require targeted detection, direct binding, and functional validation. Full article

Population ageing and the increasing digitalization of urban mobility have created new challenges for navigation services for older adults. Existing augmented reality (AR) navigation research has largely focused on interface interaction and wayfinding, with less attention focused on coordination across the wider travel service process. This study adopts a service design perspective to develop a conceptual age-friendly AR navigation product-service system. User research with 196 older adults in Nanjing, China, combined questionnaire findings with user portrait, journey, and stakeholder analyses. Eight requirement dimensions were prioritized using AHP and entropy weighting. QFD related 22 lower-level requirements to 15 functional elements, followed by PCN analysis of service activities across urban travel stages. Safety, operational ease of use, cognitive support, and sensory support emerged as the leading priorities, while traffic resource integration, AR road-condition display, and route indication ranked highest among the functional elements. The service analysis also identified opportunities to reorganize route planning, navigation assistance, transport coordination, and payment activities. The resulting conceptual system links wearable AR navigation, mobile functions, and backend service coordination, offering a structured approach to age-friendly navigation design. Full article

Training under extreme class imbalance (>1:100) remains an open problem in weakly supervised learning. The standard remedy—loss-level reweighting (focal loss, asymmetric loss, class-balanced loss)—is widely adopted, yet its behavior at extreme ratios in Multiple Instance Learning (MIL) is poorly understood. On digital breast tomosynthesis (attention-based pooling over frozen EfficientNet-B3 features), we study the optimization bounds under extreme bag-level MIL imbalance (1:251), intervening at two levels: the loss surface (via reweighting) and the gradient dynamics (via a novel q-calculus gradient modification using the Jackson q-derivative). All three reweighting strategies degrade classification relative to unweighted binary cross-entropy (BCE), monotonically, eliminating the loss surface as the bottleneck. Extended evaluation ($n=20$ seeds) shows q-calculus gradient smoothing matches vanilla BCE ($p=0.632$, Cohen’s $d=0.003$) despite provably reducing gradient variance, establishing an empirical ceiling on the optimization-achievable area under the precision-recall curve (AUPRC) of 0.0912; loss reweighting defines the floor at 0.055. Focal loss is additionally catastrophically miscalibrated (ECE > 0.44 vs. 0.036 for vanilla BCE), a collapse that persists under adaptive binning. In this regime, exceeding the ceiling points to the data-representation level, not the optimizer. We further identify ratio-invariant safety—non-degradation at any imbalance ratio, satisfied by vanilla BCE and q-calculus but violated by all reweighting methods—and give recommendations spanning moderate to extreme imbalance. Full article