Scholarlink

<p>In this paper, a real-time awareness based model for adaptive detection of driver drowsiness using fuzzy logic approach has been proposed. Using a Mamdani’s-type fuzzy inference system (FIS), 14 main predictors in four categories (time-decision, criticality, eye metrics, and driver experience) are used to assess drowsiness levels. The real-time dataset is used to normalize things like pupil size, decision response times, and driving frequency. This modular fuzzy system cascades inference outputs to assess subtle distinctions in driver states and environmental dynamics. This study validates that fuzzy logic model faithfully preserves predictive accuracy compared to machine learning models, while also reducing computational efforts with improved interpretability. Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) metrics show the robustness of the system for the applications in the real-time. In addition, it brings an efficient and interpretable framework for adaptive monitoring, pushing forward driver assistance technologies with valuable safety advantages in alleviating fatigue-related hazards.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Dynamic fuzzy logic model for adaptive driver drowsiness detection systems

  • 期刊论文

  • 发表日期：

    2025年7月25日

查看全文
  • Mallareddy Adudhodla,
  • K Prasada Rao,
  • Valiveti Dattatreya,
  • Vasavi Bande,
  • T. Prabhakara Rao,
  • Yerragudipadu Subbarayudu

摘要

In this paper, a real-time awareness based model for adaptive detection of driver drowsiness using fuzzy logic approach has been proposed. Using a Mamdani’s-type fuzzy inference system (FIS), 14 main predictors in four categories (time-decision, criticality, eye metrics, and driver experience) are used to assess drowsiness levels. The real-time dataset is used to normalize things like pupil size, decision response times, and driving frequency. This modular fuzzy system cascades inference outputs to assess subtle distinctions in driver states and environmental dynamics. This study validates that fuzzy logic model faithfully preserves predictive accuracy compared to machine learning models, while also reducing computational efforts with improved interpretability. Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) metrics show the robustness of the system for the applications in the real-time. In addition, it brings an efficient and interpretable framework for adaptive monitoring, pushing forward driver assistance technologies with valuable safety advantages in alleviating fatigue-related hazards.