Author Login IJIRT Partner Network (IPN) Login

Real-Time implementation of IoT with GCRNN and Fine-Tuning via Ninja Optimizer for Human Activity Recognition-Based Fall Detection for hospitalized patients

  • Unique Paper ID: 181718
  • Volume: 12
  • Issue: 1
  • PageNo: 5157-5166
  • Abstract:
  • Human Activity Recognition (HAR) based research develops innovative frameworks to enhance recognition performance. In this research topic utilizes a Graph Convolutional Recurrent Neural Network (GCRNN) to effectively capture both spatial and temporal dependencies in motion data collected from wearable sensors. The GCRNN model integrates Graph Convolutional Networks (GCN) to learn spatial relationships among sensor nodes, while recurrent layers model sequential dependencies, enhancing the classification of fall-related activities. To further optimize model performance, we incorporate the Ninja Optimizer, a novel optimization technique designed to improve learning rate adaptation and parameter fine-tuning. The Ninja Optimizer accelerates convergence, mitigates overfitting, and enhances generalization, leading to more reliable fall detection outcomes. We conduct experiments on the KFall dataset, demonstrating that our GCRNN model, fine-tuned with the Ninja Optimizer, outperforms conventional machine learning classifiers and baseline deep learning models. The model is first trained using the dataset and later tested with real-time sensor values. After getting the dataset, we used preprocessing of the KFall dataset, it involves multiple steps to ensure the data is clean and suitable for training the GCRNN model. First, the dataset is loaded using Pandas, and missing values are handled through mean imputation. Next, sensor data from Sensor, are normalized using Min-Max Scaling to bring all features into a uniform range. Our system employs Accelerometers, Gyroscopes, and Heart Rate Sensors to predict fall events accurately. The Heart Rate Sensor is used to detect sudden heart rate fluctuations after a fall, helping differentiate real falls from normal activities. It enhances fall detection accuracy by providing physiological insights, enabling timely medical intervention for hospitalized patients. Additionally, IoT integration enables real-time transmission of sensor data from patients to the hospital control room, ensuring continuous monitoring of patient activities. The results of this study highlight the potential of GCRNN-based Human Activity Recognition (HAR) models in improving fall detection accuracy and reliability for hospitalized patients as accuracy of 98. 93% respectively.
Download article email to a friend

Cite This Article

  • ISSN: 2349-6002
  • Volume: 12
  • Issue: 1
  • PageNo: 5157-5166

Real-Time implementation of IoT with GCRNN and Fine-Tuning via Ninja Optimizer for Human Activity Recognition-Based Fall Detection for hospitalized patients

Related Articles

Impact Factor
8.01 (Year 2024)
UGC Approved
Journal no 47859

Join Our IPN

IJIRT Partner Network

Submit your research paper and those of your network (friends, colleagues, or peers) through your IPN account, and receive 800 INR for each paper that gets published.

Join Now

Latest Publication

Archive

View More

Recent Conferences

NCSEM 2024

National Conference on Sustainable Engineering and Management - 2024 Last Date: 15th March 2024

Submit inquiry