Smart Chair for Monitoring of Sitting Behavior

Mengjie Huang, I. M. Gibson, Rui YangKnE 2017 · DOI

Abstract

Sitting is a common behavior of human body in daily life. It is found that poor sitting postures can link to pains and other complications for people in literature. In order to avoid the adverse effects of poor sitting behavior, we have developed a highly practical design of smart chair system in this paper, which is able to monitor the sitting behavior of human body accurately and non-invasively. The pressure patterns of eight standardized sitting postures of human subjects were acquired and transmitted to the computer for the automatic sitting posture recognition with the application of artificial neural network classifier. The experimental results showed that it can recognize eight sitting postures of human subjects with high accuracy. The sitting posture monitoring in the developed smart chair system can help or promote people to achieve and maintain healthy sitting behavior, and prevent or reduce the chronic disease caused by poor sitting behavior. These promising results suggested that the presented system is feasible for sitting behavior monitoring, which can find applications in many areas including healthcare services, human-computer interactions and intelligent environment.

Introduction

In modern life, individuals spend a significant portion of their day sitting—whether in office chairs, car seats, or at home. Unfortunately, poor sitting posture is a leading contributor to lumbar flexion and high compressive forces in the spine, which eventually manifest as chronic pain and long-term musculoskeletal disorders.

While researchers have previously explored posture monitoring, many existing solutions are either too invasive, too expensive, or computationally heavy. This paper presents a "smart chair" that uses a thin, flexible pressure sensor array and an Artificial Neural Network (ANN) to identify common sitting postures with high accuracy. This non-invasive system is designed to provide real-time feedback, helping users correct poor sitting habits before they lead to chronic health issues.

System Architecture and Methodology

The system architecture is divided into two main components: hardware sensing and software classification.

Hardware Setup

The team utilized a piezo-resistive sensor array, a material where electrical resistance changes based on applied pressure. While the array was capable of a high-resolution 52x44 grid, the researchers opted for a sparse 8x8 grid (64 sensing units) to maximize efficiency and reduce cost. The sensor is remarkably thin (0.25mm), making it virtually unnoticeable to the user when integrated into a standard chair seat.

Figure 1: The pressure sensor array layout (left) and the circuit diagram for the individual sensing units (right).

Software and Classification

To ensure the system works for people of different weights, raw pressure data was normalized to an interval of [-1, 1]. This allows the algorithm to focus on the pattern of the pressure distribution rather than the absolute weight of the individual. The heart of the system is an Artificial Neural Network (ANN) implemented in MATLAB, featuring 20 hidden neurons using a logarithmic sigmoid (logsig) transfer function and a Scaled Conjugate Gradient (SCG) backpropagation training method.

Experimental Setup and Dataset

The researchers generated a custom dataset to train and validate the system, focusing on eight standardized sitting positions common in daily life.

Figure 2: The eight standardized sitting postures (P1-P8) used to train and test the smart chair system.
  • Postures: Upright (P1), Slumped (P2), Leaning Forward (P3), Leaning Backward (P4), Leaning Left (P5), Leaning Right (P6), Right Leg Crossed (P7), and Left Leg Crossed (P8).
  • Sample Size: 40 sets of pressure data per posture were collected for training, with another 40 sets per posture reserved for verification.
  • Protocol: Each posture was held for 5 seconds across 5 trials per subject to ensure the body reached a state of equilibrium before data capture.

Key Results and Insights

The ANN classifier achieved an overall accuracy of 92.2%. The study yielded several critical insights into the feasibility of sparse pressure sensing:

  • Efficiency of Sparse Sensing: A simple 8x8 grid provided sufficient spatial information for the ANN to distinguish between complex positions, proving that high-resolution maps are not necessary for accurate posture recognition.
  • Posture-Specific Performance: Performance varied by posture. Leaning Left, Leaning Right, and Left Leg Crossed were identified with near-perfect accuracy. However, Slumped sitting (P2) was the most difficult to identify (72.7% accuracy), as it was frequently confused with upright or leaning backward positions due to the hip remaining relatively fixed in the center of the seat.
  • Computational Speed: The model reached its performance goal (a mean square error of less than 0.01) in just 166 training epochs, suggesting the system is computationally light enough for real-time monitoring applications.

Conclusion and Future Work

The presented smart chair system demonstrates a feasible, non-invasive path toward monitoring sitting behavior in real-time. By normalizing data, the researchers have moved toward a "one-size-fits-all" software solution that does not require recalibration for different body weights.

Future work will focus on transitioning from static to dynamic sitting behavior to account for fidgeting and movement. Additionally, the researchers plan to incorporate specific "hip location" tracking to better differentiate between subtle shifts in the pelvis, which should improve the accuracy of distinguishing between upright, slumped, and leaning backward positions.

Summary updated

Abstract

<p class="1">Sitting is a common behavior of human body in daily life. It is found that poor sitting postures can link to pains and other complications for people in literature. In order to avoid the adverse effects of poor sitting behavior, we have developed a highly practical design of smart chair system in this paper, which is able to monitor the sitting behavior of human body accurately and non-invasively. The pressure patterns of eight standardized sitting postures of human subjects were acquired and transmitted to the computer for the automatic sitting posture recognition with the application of artificial neural network classifier. The experimental results showed that it can recognize eight sitting postures of human subjects with high accuracy. The sitting posture monitoring in the developed smart chair system can help or promote people to achieve and maintain healthy sitting behavior, and prevent or reduce the chronic disease caused by poor sitting behavior. These promising results suggested that the presented system is feasible for sitting behavior monitoring, which can find applications in many areas including healthcare services, human-computer interactions and intelligent environment.</p>
BibTeX
@inproceedings{huang2017smart,
  author = {Mengjie Huang and I. M. Gibson and Rui Yang},
  title = {{Smart Chair for Monitoring of Sitting Behavior}},
  booktitle = {KnE},
  year = {2017},
  doi = {10.18502/keg.v2i2.626},
}
Figure