Development of a mixed reality-based learning system with dynamic posture recognition for advancing ergonomics education
No Thumbnail Available
Authors
Meeting name
Sponsors
Date
Journal Title
Format
Thesis
Subject
Abstract
Proper lifting posture is one of the crucial topics in ergonomics education. Prolonged use of incorrect postures to carry heavy objects can lead to potential illnesses. The high rate of lumbar spine injuries among airport baggage handlers is an example of this. To enhance students' understanding, we designed a series of mixed reality modules enabling them to view the working process of airport baggage handlers and learn the biomechanical calculations related to the lifting task through Microsoft HoloLens 2. In the last module, students are required to perform the baggage handling movement with Xsens MVN Awinda motion capture sensors worn, and the system provides feedback based on what they performed. We recruited 20 undergraduate and graduate students to try out this learning system, including 10 males and 10 females. We evaluated this system from multiple aspects. In terms of system performance, there were no application errors or data loss during the experiments, which demonstrated the robustness of our system architecture and implementation. We obtained 39 recognition results with a recognition accuracy of 92.31 percent. We tested with different baseline movements on the recorded participants' motion data after the experiments but found that the accuracy was not improved. This suggests the limitations of our method in classifying similar movements with small differences. In terms of learning performance, the 20 participants watched lectures introducing proper lifting postures but 9 of them still chose to perform the baggage handling movement with a bad posture. This illustrates the limitations of the instructional effectiveness of listening to lectures and highlights the significance of hands-on learning. The participants' SUS scores demonstrate that our system usability is between okay and good. Since the participants' SCLS results indicate that our educational materials are helpful and their NASA TLX results indicate that the tasks we designed are easier to complete, we can improve the system usability by providing more interaction in the virtual environment or designing additional engaging learning activities. In conclusion, we developed a robust mixed reality-based learning system with dynamic posture recognition capabilities, which benefits ergonomics education. Our proposed framework is promising to be applied to more learning scenarios in which students follow specific instructions.
Table of Contents
PubMed ID
Degree
M.S.