SitR: Sitting Posture Recognition Using RF Signals

Lin Feng,Ziyi Li,Chen Liu,Xiaojiang Chen,Xiao Yin,Dingyi Fang
DOI: https://doi.org/10.1109/jiot.2020.3019280
IF: 10.6
2020-12-01
IEEE Internet of Things Journal
Abstract:Sitting posture has a close relationship with our health, and keeping a healthy sitting posture is critical to each of us. Poor sitting postures often inevitably increase the risk of modern health musculoskeletal disorders. Previous works either used a camera to record the image or attached wearable sensors on the human body to recognize sitting postures. However, video-based approaches may face privacy issues while the wearable sensor-based approaches may cause uncomfortable to the user. In this work, we propose SitR, the first sitting posture recognition system using RF signals alone, which neither compromises the privacy nor requires wearing various sensors on the human body. We demonstrate that SitR can successfully recognize seven habitual sitting postures with just three lightweight and low-cost RFID tags pasted to the user's back. Our design exploits the correlation between the phase change of RFID tags and the sitting postures. By extracting effective features of the measured phase sequences and employing appropriate machine learning algorithm, SitR can achieve robust and high performance. We evaluate the performance of SitR with ten volunteers in two different scenarios. Extensive experiments show SitR can recognize seven sitting postures with a high accuracy across different scenarios and various conditions. SitR can further detect the abnormal respiration, stand up, and sit down and provide sitting posture history for sedentary people.
computer science, information systems,telecommunications,engineering, electrical & electronic
What problem does this paper attempt to address?
The paper mainly aims to address the following issues: ### Research Background and Objectives With changes in modern lifestyles, people often maintain poor sitting postures for extended periods, which can lead to various health problems such as cervical spondylosis and chronic back pain, among other musculoskeletal disorders. Therefore, developing a method that can monitor sitting posture without invading privacy and without causing inconvenience to users has become particularly important. ### Solution The paper proposes a new system called SitR, which uses Radio Frequency Identification (RFID) technology to identify seven common sitting postures, including sitting upright, leaning forward, leaning backward, resting the face on the hand (left or right hand), and crossing legs (left leg or right leg on top). The system only requires three lightweight and low-cost RFID tags to be attached to the user's back to work, without the need for cameras or other wearable devices. ### Key Technologies and Contributions 1. **Deployment Strategy**: The study found that attaching the tags to the back can effectively capture the information needed to distinguish different sitting postures without causing discomfort to the user. Experiments verified the impact of different numbers of tags on recognition accuracy, and three tags were ultimately chosen as the optimal configuration. 2. **Data Preprocessing**: - **Phase Offset Calibration Algorithm**: Phase offsets caused by hardware issues can affect recognition accuracy. The paper proposes an algorithm to detect and correct these phase offsets. - **Activity Detection Algorithm**: Used to detect activities such as "standing up" and "sitting down." 3. **Feature Extraction**: - Utilizes features from both the time domain and frequency domain for posture recognition. Effective features are extracted by analyzing the phase sequence changes of RFID tags under different sitting postures. 4. **Experimental Evaluation**: - Tests in real-world environments show that SitR can recognize seven sitting postures with an average accuracy of 98.83% under different scenarios and conditions. - Additionally, the paper evaluates the impact of different antenna gains, chair materials, and clothing thickness on system performance, demonstrating the system's robustness. ### Conclusion SitR is an innovative solution that can achieve efficient and accurate sitting posture recognition through a simple RFID tag configuration without invading privacy. This has significant implications for promoting healthy habits among people.