Falls are a significant health concern globally, especially among the elderly. Traditional fall detection systems, while useful, often suffer from being costly, intrusive, and inaccurate. Our project introduces a novel approach to fall detection using Wi-Fi sensing, a non-invasive method leveraging existing wireless network infrastructure. The core of our project is a signal processing framework designed specifically for Wi-Fi based fall detection. This framework, developed through extensive research and testing, utilizes advanced filtering, feature extraction, and predictive algorithms to enhance detection accuracy and efficiency. A key achievement of our framework is its superior performance over previous systems, as validated by a Multivariate analysis of variance (MANOVA), which showed a statistically significant improvement with a p-value of 0.026 (alpha = 0.05). This framework not only marks a significant advancement in fall detection technology but also lays the groundwork for future innovations in Wi-Fi sensing applications.
To install the fall detection framework, follow these steps:
-
System Requirements: Ensure your system meets the following requirements:
- Operating System: Windows 10 or later, MacOS 10.14 or later, Linux distribution supporting Python 3.6+
-
Installing Dependencies:
- Install Python 3.6 or later from Python's official website.
- Install necessary Python libraries using pip:
pip install numpy scipy sklearn pandas
-
Downloading the Software:
- Clone the repository from GitHub:
git clone https://github.com/your-github-repo/fall-detection-wifi.git
- Navigate to the cloned directory:
cd fall-detection-wifi
- Clone the repository from GitHub:
To use the fall detection system, run each file with the UT-HAR dataset. Extract each set of data features, for each data filter, and pass the data to each machine learning algorithm for evaluvation.
This project is licensed under the MIT License. This means that it is free for personal and commercial use, modification, and distribution with attribution. For full license details, refer to the LICENSE file in the repository.
I would like to thank the following organizations and individuals for their assistance in completing this research project. First, the Monmouth County Board of Commissioners for their continuing support of MCVSD and the work they continue to do for our district. Secondly, the MCVSD Board of Education and Administration for providing the MCVSD and High Technology High School the resources that enable the high school research program to continue. Next, all the administrators and faculty of High Technology High School for providing me the opportunity to pursue and research my field and who continue to support me throughout my journey of growing as a researcher. All the research teachers and mentors who have taught me valuable research skills which enabled my progress to be a success. Lastly, my parents for their continuing support and finances that helped this project come to fruition.