Device-Free Human Activity Recognition Using Commercial WiFi Devices

By Wen Wang, Alex X. Liu, Muhammad Shahzad, Kang Ling, Sanglu Lu. IEEE Journal on Selected Areas in Communications, 2017

465 citations recorded by the source corpus.

Since human bodies are good reflectors of wireless signals, human activities can be recognized by monitoring changes in WiFi signals. However, existing WiFi-based human activity recognition systems do not build models that can quantify the correlation between WiFi signal dynamics and human activities. In this paper, we propose a Channel State Information (CSI)-based human Activity Recognition and Monitoring system (CARM). CARM is based on two theoretical models. First, we propose a CSI-speed model that quantifies the relation between CSI dynamics and human movement speeds. Second, we propose a CSI-activity model that quantifies the relation between human movement speeds and human activities. Based on these two models, we implemented the CARM on commercial WiFi devices. Our experimental results show that the CARM achieves recognition accuracy of 96% and is robust to environmental changes.

Subfields: Computer Networks and Communications, Computer Vision and Pattern Recognition, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/f312682ae93523de41a4f1e2bc0ca30e4345c8a4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

1. Gang Zhou
2. Jun Liu
3. Kun Qian
4. Yunhao Liu
5. Hongbo Liu
6. Guidong Zhang
7. Yue Zheng
8. Jian Liu

In Computer Vision and Pattern Recognition

1. Qiuhong Ke
2. Jun Liu
3. Hossein Rahmani
4. Zehua Sun
5. Kun Qian
6. Gang Zhou
7. Yunhao Liu
8. Guidong Zhang

In Electrical and Electronic Engineering

1. Zheng Yang
2. Yunhao Liu
3. Kun Qian
4. Gang Zhou
5. Chenshu Wu
6. Guidong Zhang
7. Yi Zhang
8. Yongsen Ma

Citations in the corpus

• Wireless Sensing for Human Activity: A Survey (2020, methodology): “[22], [23] propose a human activity recognition system, named CARM, which builds CSI-speed model and CSI-activity model to quantify the correlation between the movement speeds of different human body parts and a specific human activity.”
• Cross-Domain WiFi Sensing with Channel State Information: A Survey (2022, background): “On the other hand, for most recognition applications, such as activity recognition [88, 91], gesture recognition [1, 47], and human/object identification [8, 9], the number of classes that need to be recognized ranges from two to several hundred.”
• Widar3.0: Zero-Effort Cross-Domain Gesture Recognition With Wi-Fi (2021, methodology): “Given that the range of human-induced DFS is within (cid:5) 60 Hz [11], the DFS profile of one link can only provide about 12 constraints.”
• Zero-Effort Cross-Domain Gesture Recognition with Wi-Fi (2019, methodology): “, power profiles of Doppler frequency shifts [44]) from Wi-Fi signals and map them to human gestures.”
• WiFi CSI Based Passive Human Activity Recognition Using Attention Based BLSTM (2019, methodology): “Therefore, many research works did not include this information for recognition [13], [14].”
• WiFi Sensing with Channel State Information (2019, background): “CARM [98, 99] NR: CFO; ST: DWT; SE: Thresholding, PCA, Feature Extraction L: HMM Activity Recognition Recognition Accuracy: >96% (8 activities)”
• Human Action Recognition From Various Data Modalities: A Review (2020, methodology): “Specifically, most of the existing WiFi-based HAR methods [292], [293] focused on using the Channel State Information (CSI) to conduct the HAR task.”