KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/kmitl/1

Browse

Settings

Access type: UnknownSubject: search.filters.subject.Bayesian optimization

Search Results

Now showing 1 - 2 of 2
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Deep learning-based comparative evaluation of EEG, HRV, and EDA biomarkers for personal thermal comfort prediction
    (2026-03-01)
    Sahoh, Bukhoree
    ;
    Wongsontham, Fatimah
    ;
    Tipsavak, Apaporn
    ;
    Chaithong, Paweena
    ;
    Kliangkhlao, Mallika
    Personal thermal comfort prediction significantly impacts health, well-being, and productivity, yet existing systems typically employ multiple physiological biomarkers without clear evidence for optimal single-modality solutions. This study presents a deep learning (DL)-based metrologically grounded framework to evaluate and compare three biomarkers—electroencephalography (EEG), heart rate variability (HRV), and electrodermal activity (EDA)—as traceable sensors under precisely controlled thermal conditions. We developed specialized signal preprocessing and feature engineering pipelines for each modality: 1) EEG spectral decomposition via Welch's power spectral density estimation across frequency bands (delta, theta, alpha, beta, and gamma); 2) HRV analysis through Fast Fourier Transform spectral estimation focusing on low-frequency to high-frequency component ratios; and 3) EDA feature extraction utilizing optimized filtering techniques to isolate skin conductance level and response characteristics. Three biomarker-specific DL architectures undergo Bayesian hyperparameter optimization, enabling equitable comparison of each modality's predictive performance. Results demonstrate that the HRV-based model achieves superior performance (F-measure = 0.95) while requiring minimal computational resources (0.78 MB memory footprint, 0.89 relative cost compared to the EEG baseline), establishing it as the optimal single-modality solution. These findings provide a paradigm for reproducible physiological measurement systems, advancing thermal comfort prediction by combining clinical-grade reliability with practical implementation benefits for clinical applications and next-generation indoor environmental control systems.
      1
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Novel Adaptive Intelligent Control System Design
    (2025-08-01)
    Duanyai, Worrawat
    ;
    Song, Weon Keun
    ;
    Ka, Min Ho
    ;
    Lee, Dong Wook
    ;
    Dissanayaka, Supun
    A novel adaptive intelligent control system (AICS) with learning-while-controlling capability is developed for a highly nonlinear single-input single-output plant by redesigning the conventional model reference adaptive control (MRAC) framework, originally based on first-order Lyapunov stability, and employing customized neural networks. The AICS is designed with a simple structure, consisting of two main subsystems: a meta-learning-triggered mechanism-based physics-informed neural network (MLTM-PINN) for plant identification and a self-tuning neural network controller (STNNC). This structure, featuring the triggered mechanism, facilitates a balance between high controllability and control efficiency. The MLTM-PINN incorporates the following: (I) a single self-supervised physics-informed neural network (PINN) without the need for labelled data, enabling online learning in control; (II) a meta-learning-triggered mechanism to ensure consistent control performance; (III) transfer learning combined with meta-learning for finely tailored initialization and quick adaptation to input changes. To resolve the conflict between streamlining the AICS’s structure and enhancing its controllability, the STNNC functionally integrates the nonlinear controller and adaptation laws from the MRAC system. Three STNNC design scenarios are tested with transfer learning and/or hyperparameter optimization (HPO) using a Gaussian process tailored for Bayesian optimization (GP-BO): (scenario 1) applying transfer learning in the absence of the HPO; (scenario 2) optimizing a learning rate in combination with transfer learning; and (scenario 3) optimizing both a learning rate and the number of neurons in hidden layers without applying transfer learning. Unlike scenario 1, no quick adaptation effect in the MLTM-PINN is observed in the other scenarios, as these struggle with the issue of dynamic input evolution due to the HPO-based STNNC design. Scenario 2 demonstrates the best synergy in controllability (best control response) and efficiency (minimal activation frequency of meta-learning and fewer trials for the HPO) in control.
      1