Human Activity Recognition Using CSI Information with Nexmon

Jörg Schäfer,Jens Liebehenschel,Hannan Adil,Baldev Raj Barrsiwal,Muyassar Kokhkharova

doi:10.3390/app11198860

Abstract

Using Wi-Fi IEEE 802.11 standard, radio frequency waves are mainly used for communication on various devices such as mobile phones, laptops, and smart televisions. Apart from communication applications, the recent research in wireless technology has turned Wi-Fi into other exploration possibilities such as human activity recognition (HAR). HAR is a field of study that aims to predict motion and movement made by a person or even several people. There are numerous possibilities to use the Wi-Fi-based HAR solution for human-centric applications in intelligent surveillance, such as human fall detection in the health care sector or for elderly people nursing homes, smart homes for temperature control, a light control application, and motion detection applications. This paper’s focal point is to classify human activities such as EMPTY, LYING, SIT, SIT-DOWN, STAND, STAND-UP, WALK, and FALL with deep neural networks, such as long-term short memory (LSTM) and support vector machines (SVM). Special care was taken to address practical issues such as using available commodity hardware. Therefore, the open-source tool Nexmon was used for the channel state information (CSI) extraction on inexpensive hardware (Raspberry Pi 3B+, Pi 4B, and Asus RT-AC86U routers). We conducted three different types of experiments using different algorithms, which all demonstrated a similar accuracy in prediction for HAR with an accuracy between 97% and 99.7% (Raspberry Pi) and 96.2% and 100% (Asus RT-AC86U), for the best models, which is superior to previously published results. We also provide the acquired datasets and disclose details about the experimental setups.

Highlights

One of the most discussed recent research topics in wireless technology and smart home applications is human activity recognition (HAR), and there are numerous applications such as health care, ambient assisted living, and children and elderly people monitoring systems.In 2017 in 22 Western European countries there were reported at least 8.4 million injury cases due to falls
In our work which is mainly based on the master theses of some of the authors, we explore channel state information (CSI) data acquisition possibilities using Nexmon with a Raspberry Pi 3 B+, Pi 4B [13], and Asus RT-AC86U based platform [14,15] and human activities classification using machine learning algorithms with a focus on practical deployment
Multiclass support vector machine (SVM) with the One-to-One approach and with polynomial kernel and errorcorrecting output codes (ECOC) model is applied to classify the four activities, SIT, STAND, WALK, and EMPTY by using the features mentioned in Section 7.5 for training the network

Summary

Introduction

One of the most discussed recent research topics in wireless technology and smart home applications is human activity recognition (HAR), and there are numerous applications such as health care, ambient assisted living, and children and elderly people monitoring systems.In 2017 in 22 Western European countries there were reported at least 8.4 million injury cases due to falls. One of the most discussed recent research topics in wireless technology and smart home applications is human activity recognition (HAR), and there are numerous applications such as health care, ambient assisted living, and children and elderly people monitoring systems. A total of 5667 per 100,000 people in the age range 70–74 years and 47,239 per 100,000 people in the age category of 95+ years required medical help due to injuries. These numbers are greatly underestimated, since not every elderly person reports every fall. HAR for motion detection and human fall detection could benefit the medical staff in hospitals, elderly people care houses and people with paralysis, epilepsy, hypoglycemia diabetic diseases, and others. Recent advancements in Wi-Fi-based machine learning solutions could become an alternative and simplified solution for usability and personal privacy concerns [3]

Methods

Results

Conclusion