Real-Time Contactless WiFi Based Room Detection of Sitting and Standing Human Motions

In the field of healthcare, Human activity monitoring has recently been gaining widespread attention. The ability to monitor human activities is being applied to assist the care of vulnerable people. These monitoring systems can allow elderly people to live more independent lives within their own homes without residing in care facilities. The implementation of monitoring systems can relieve the strain on family members and/or caregivers from frequent visits to check on vulnerable people's well-being. The work of this paper proposes a contactless real-time monitoring system to detect if a person is sitting or standing. The contactless feature works by using machine learning to classify the propagation of RF signals as they travel through the atmosphere. The propagation data is collected using local devices and then uploaded to the cloud. A dashboard is used to download the data from the cloud and provide information on the output of the monitoring system. The system uses data filtering techniques to observe the patterns of propagation and establish if movements have taken place. If movements are detected then the data is passed to a trained AI model. The AI model will classify the movements as Sitting or Standing. The training of the AI model included applying 10-fold cross-validation to the training data to test performance. The Random Forest algorithm achieved an accuracy of 90.75 % and was used to build the AI model.