Smart bodysuit integrating digital twin technology for real-time human motion monitoring and visualization

Abstract

Real-time motion monitoring can have many significant applications in fields such as smart healthcare, humancomputer interaction, sports training, and fashion metaverse. However, current wearable systems used for motion monitoring suffer from issues related to usability, comfort, interactivity, and accuracy. To address these challenges, this work proposes a smart bodysuit system that integrates digital twin (DT) technology for real-time human motion monitoring and visualization. The system comprises novel hardware and software, including a specially designed knit bodysuit housing the electronic components to reduce calibration errors, a data calibration method to take into account the possible clothing-skeleton misalignment, an adaptive-threshold method combining random forest (RF) and zero-speed update algorithms to enhance motion recognition accuracy, and an enhanced TCP-Illinois congestion control algorithm to reduce system latency. Furthermore, DT technology is incorporated to improve humancomputer interaction. Wearer trial tests showed that the system exhibits low latency, high accuracy, and good comfort and wearability. The network latency test showed a delay of 13.17 ms, which is comparable to or even better than commercial systems. The testing error of the joint angle can be controlled within 0.2°, and the recognition accuracy of the motion pattern can reach 98.6%. The proposed adaptive-threshold method has higher positioning accuracy compared to the existing fixed-threshold method. Furthermore, the smart bodysuit system can reconstruct the real-time wearer’s motion states through DT technology. The developed system is believed to have potentially significant applications in the metaverse, digital healthcare, home monitoring, and sports science.