欢迎访问王栋教授课题组
欢迎具有材料、化学、纺织、环境、机械相关背景的研究生、青年教师加入本课题组
设为首页 | 收藏本站

2024-01-03

2024-01-03 16:10浏览数:102

Accurate and Efficient Sitting Posture Recognition and Human-Machine Interaction Device Based on Fabric Pressure Sensor Array and Neural Network

Abstract: This article presents a novel approach to interact with users through posture recognition, leveraging its advantages of convenience and real-time feedback to enhance user engagement and personalized experiences. In contrast to traditional methods that rely on camera-based posture detection, this study proposes a deep learning-based framework for posture recognition by classifying the distribution of body pressure under different sitting positions. The system integrates a large-area, highly flexible fabric pressure sensor array into the chair, which collects data on posture-specific pressure patterns for training and identification purposes. A deep learning algorithm, specifically the LeNet architecture, is employed to classify 49 different sitting positions based on angular variations, including body tilt to the left or right, standard posture, and forward or backward leaning. The proposed approach achieves an impressive accuracy rate of 99.86%. Furthermore, the application of this posture recognition system in VR devices enables intelligent chair control for VR games. This research provides strong support for future advancements in chair design and human-computer interaction technologies, enhancing ergonomic designs in various domains such as automotive seats, office chairs, and medical seating, while simultaneously improving user comfort and well-being.

20240207_10.1002_admt.202301579.jpg

Weibing Zhong, Hui Xu, Yiming Ke, Xiaojuan Ming, Haiqing Jiang, Mufang Li, Dong Wang. Accurate and Efficient Sitting Posture Recognition and Human-Machine Interaction Device Based on Fabric Pressure Sensor Array and Neural Network. Advanced Materials Technologies. 2024; 9(3): 2301579.


会员登录
 
 
登录