Kou Yamada

This paper proposes a federated learning framework for privacy-preserving patient monitoring with lightweight blockchain anchoring. The framework keeps synthetic patient monitoring records local at each client and uses federated model aggregation to support collaborative learning without centralizing raw records. To improve traceability, the blockchain layer is specified as an anchoring mechanism that records compact evidence, including model hashes and participation metadata, rather than raw data or full model parameters. Experiments were conducted on synthetic patient monitoring data to evaluate framework behavior under non-IID client distributions, label noise, different client counts, partial client participation, and aggregation strategies. The centralized MLP baseline achieved approximately 0.89 overall accuracy but failed to detect alert cases, with 0% alert-class recall, showing that accuracy alone can be misleading in imbalanced monitoring scenarios. In the federated simulations, the model reached approximately 0.99 accuracy under clean labels, approximately 0.90 under 10% label noise, and approximately 0.70 under 30% label noise. Under a more difficult noisy, non-IID, dropout, and fixed skewed-client evaluation setting, the model stabilized at approximately 0.80 accuracy after 25 communication rounds. Client scaling from 5 to 20 clients remained stable, and FedAvg, weighted aggregation, and accuracy-trimmed robust aggregation produced similar final accuracy of approximately 0.98 in the 10-client setting. The results indicate that label quality strongly affects federated convergence, while blockchain anchoring should be interpreted as an auditability mechanism rather than a direct accuracy-improving component. This study provides a framework-level foundation for auditable federated patient monitoring in semi-trusted healthcare networks.

Real-time traffic safety analysis is necessary to determine proactive risks associated with traffic conflicts. This study presents a driving-risk detection framework based on high-precision naturalistic vehicle trajectory data to support host vehicle driver-warning systems. A rear-end conflict identification approach based on time-to-collision (TTC) is employed to detect vehicle interactions as safe or unsafe. To capture temporal driving patterns, frame-level observations are transformed into sequential samples using a sliding-window strategy while preserving the natural class imbalance of real-world traffic data. Several conventional machine learning models, including CatBoost, LightGBM, XGBoost, Random Forest, Extra Trees, Decision Tree, and SVM, as well as recurrent deep learning models such as Simple RNN, LSTM, and GRU, are evaluated using leave-one-subject-out cross-validation across seven expressways. Among the evaluated models, the Simple RNN achieves a recall of 99.12% and an F1-score of 98.48%, outperforming the conventional machine learning models. Its predictive performance is comparable to that of LSTM and GRU while offering lower inference latency, making it suitable for real-time deployment. A SHAP-based Explainable AI analysis is conducted to identify the most influential factors in conflict detection and to provide insight into model predictions. The analysis supports that host vehicle speed, preceding vehicle speed, and inter-vehicle gap are the primary determinants of rear-end conflict risk. This proactive and interpretable framework, evaluated offline on naturalistic trajectory data, demonstrates strong potential for integration into real-time driver-warning systems.

Unregulated parking search is a major source of urban inefficiency, leading to unnecessary cruising, wasted time, driver frustration and increased emissions. Moreover, when drivers select spots solely based on immediate self‐interest, ignoring others' dynamic needs and preferences, this behaviour exacerbates congestion, conflict and overall discomfort. To address these challenges and advance towards a smart, sustainable and user‐centric mobility system, this paper proposes an Optimal Parking Allocation Scheme (OPAS). By intelligently integrating user behaviour and real‐time demand, OPAS dynamically assigns parking spaces that enhance overall system efficiency while flexibly providing each user with an optimal spot, balancing collective sustainability with individual comfort. The proposed OPAS wirelessly receives the user information upon arrival at the entrance and determines the best‐suited spot by solving an adaptive optimisation problem considering the travel and parking behaviour in terms of the number of passengers in the car, their walking and driving distances, the priority level of users, in addition to trends of the occupation ratio. The proposed OPAS dynamically tunes its objective parameters to align with overarching system goals, yielding successive, situation‐adaptive decisions. The performance of the proposed scheme is evaluated using data from an indoor parking model of a shopping centre in Gunma City, Japan. The results demonstrate that OPAS achieves a significant reduction in both individual walking distance and vehicular travel distance within the parking facility. Consequently, this efficiency directly translates to a decrease in total fuel consumption and associated emissions. Specifically, the proposed strategy achieves a daily reduction of 23 kg in emissions and 10 L in fuel consumption, highlighting the practical significance of the proposed approach.

This paper proposes an efficient parameter design approach for time-synchronized control in the spacecraft attitude stabilization problem. Existing studies on time-synchronized control typically demonstrate robustness only against external disturbances, even though most mechanical systems, including spacecraft, inherently exhibit system uncertainties such as modeling errors. Moreover, the interconnection between the control system and the disturbance observer, which is crucial for robustness, has not been clearly addressed. The main objective of this work is to design a time-synchronized controller capable of handling both model uncertainties and external disturbances while ensuring convergence faster than a desired settling time. To this end, we formulate a Linear Matrix Inequality (LMI) condition for a combined system of the time-synchronized controller and disturbance observer, guaranteeing robust finite-time stability through a rigorous Lyapunov analysis. Since the settling-time function contains a free parameter, an iterative LMI algorithm is employed to achieve a settling time as close as possible to the desired value.We further propose a tuning method based on a bisection algorithm, which improves settling-time accuracy. The proposed LMI-based approach enables optimal parameter design that ensures robustness for both the control and disturbance observer systems, while also guaranteeing convergence within the specified settling time. The effectiveness of the method is demonstrated through numerical simulations, which confirm the practical advantages of the proposed design.

Urban parking facilities often experience severe peak-period congestion, resulting in delays, fuel consumption, and emissions. This paper develops an incentive-based intelligent parking management system to address the challenges of peak demand by encouraging drivers with flexible schedules to shift their parking from peak to off-peak times. The proposed incentive model regulates peak demand, which has been calibrated using historical data on parking demand and occupancy. The model incorporates empirically derived behavioral parameters (from field surveys) to capture drivers’ sensitivity to incentives. The system’s performance is evaluated via discrete-time simulation using real-world parking data from a Japanese supermarket, considering both weekday and weekend demand patterns. The incentive mechanism redistributed approximately 6% of the total parking demand from peak to off-peak periods, markedly reducing peak congestion. This demand shift resulted in substantial sustainability benefits: CO₂ emissions decreased by approximately 21% on weekdays (19.5% on weekends), and fuel consumption decreased by about 25% on weekdays (28% on weekends) compared to a baseline scenario without incentives. The prioritizing of electric cars (EVs) and hybrid electric vehicles (HEVs) significantly enhanced emission reductions by promoting cleaner vehicles in the allocation process. This behavioral demand-management strategy offers a practical and scalable solution to enhance urban mobility and sustainability, demonstrating how modest incentives can yield substantial benefits in terms of traffic flow and emissions mitigation.

This paper presents a study on the applicability of color-based emotional expression models for robots, originally developed for Japanese users, to a Vietnamese demographic. Our research introduces a newly collected dataset capturing Vietnamese participants’ interpretations of robot-emitted colors and gradients, addressing a significant gap in culturally adapted robotic communication. Through two experimental conditions, we evaluate the effectiveness of (1) solid colors and (2) color gradients in conveying eight primary emotions. A total of 330 Vietnamese participants engaged in the study, providing subjective ratings on emotion recognition accuracy. Our findings indicate that while some emotions (e.g. anger and trust) align well with color representations, others (e.g. surprise and disgust) exhibit cultural variations in perception. A one-way ANOVA test confirms statistically significant differences in how emotions are recognized across color models. These results underscore the importance of cultural adaptation in robot emotion expression and highlight potential directions for refining color-based models for cross-cultural applications. Future research will explore neural network-based emotion adaptation to enhance real-time, culturally responsive robotic interactions.