Scholarlink

<p>Traditional security systems for women and children often fail, limited by manual activation and slow response times. This research presents an IoT-enabled smart safety framework that leverages machine learning (ML) to autonomously detect and respond to physiological distress and potential threats. The proposed system integrates physiological sensors (heart rate, temperature) and activity sensors (GPS, accelerometer) into an intelligent wearable device. A hybrid ML approach, primarily utilizing Support Vector Machine (SVM) and Naive Bayes (NB), is employed for robust activity recognition and stress level classification. Performance was rigorously validated using k-fold cross-validation, with the SVM classifier achieving a 99.7% average accuracy in threat detection. This AI-driven approach reduces detection latency to 3&#xa0;s, while battery optimization ensures 18–20&#xa0;h of continuous operation. Upon autonomous threat detection, the system uses GSM connectivity to transmit GPS coordinates to authorities. This research demonstrates a practical, high-accuracy solution for personal security, with a strong emphasis on data privacy and ethical deployment.</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

Evaluation of IoT based smart safety systems for women and children using machine learning techniques

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月29日

查看全文
  • Nanda R. Wagh,
  • Sanjay R. Sutar,
  • V. J. Kadam,
  • S. M. Jadhav,
  • A. S. Yadav,
  • V. S. Pawar

摘要

Traditional security systems for women and children often fail, limited by manual activation and slow response times. This research presents an IoT-enabled smart safety framework that leverages machine learning (ML) to autonomously detect and respond to physiological distress and potential threats. The proposed system integrates physiological sensors (heart rate, temperature) and activity sensors (GPS, accelerometer) into an intelligent wearable device. A hybrid ML approach, primarily utilizing Support Vector Machine (SVM) and Naive Bayes (NB), is employed for robust activity recognition and stress level classification. Performance was rigorously validated using k-fold cross-validation, with the SVM classifier achieving a 99.7% average accuracy in threat detection. This AI-driven approach reduces detection latency to 3 s, while battery optimization ensures 18–20 h of continuous operation. Upon autonomous threat detection, the system uses GSM connectivity to transmit GPS coordinates to authorities. This research demonstrates a practical, high-accuracy solution for personal security, with a strong emphasis on data privacy and ethical deployment.