School of Computing and Information Systems, Grand Valley State University, Allendale, MI, United States.
Computer Science, University of Nevada, Las Vegas, Las Vegas, NV, United States.
JMIR Mhealth Uhealth. 2021 Mar 5;9(3):e24194. doi: 10.2196/24194.
This study presents a new approach to measure and analyze the walking balance of humans by collecting motion sensor data in a smartphone.
We aimed to develop a mobile health (mHealth) app that can measure the walking movements of human individuals and analyze the differences in the walking movements of different individuals based on their health conditions. A smartphone's motion sensors were used to measure the walking movements and analyze the rotation matrix data by calculating the variation of each xyz rotation, which shows the variables in walking-related movement data over time.
Data were collected from 3 participants, that is, 2 healthy individuals (1 female and 1 male) and 1 male with back pain. The participant with back pain injured his back during strenuous exercise but he did not have any issues in walking. The participants wore the smartphone in the middle of their waistline (as the center of gravity) while walking. They were instructed to walk straight at their own pace in an indoor hallway of a building. The walked a distance of approximately 400 feet. They walked for 2-3 minutes in a straight line and then returned to the starting location. A rotation vector in the smartphone, calculated by the rotation matrix, was used to measure the pitch, roll, and yaw angles of the human body while walking. Each xyz-rotation vector datum was recalculated to find the variation in each participant's walking movement.
The male participant with back pain showed a diminished level of walking balance with a wider range of xyz-axis variations in the rotations compared to those of the healthy participants. The standard deviation in the xyz-axis of the male participant with back pain was larger than that of the healthy male participant. Moreover, the participant with back pain had the widest combined range of right-to-left and forward-to-backward motions. The healthy male participant showed smaller standard deviation while walking than the male participant with back pain and the female healthy participant, indicating that the healthy male participant had a well-balanced walking movement. The walking movement of the female healthy participant showed symmetry in the left-to-right (x-axis) and up-to-down (y-axis) motions in the x-y variations of rotation vectors, indicating that she had lesser bias in gait than the others.
This study shows that our mHealth app based on smartphone sensors and rotation vectors can measure the variations in the walking movements of different individuals. Further studies are needed to measure and compare walking movements by age, gender, as well as types of health problems or disease. This app can help in finding differences in gait in people with diseases that affect gait.
本研究提出了一种新的方法，通过在智能手机中收集运动传感器数据来测量和分析人类的步行平衡。
我们旨在开发一个移动健康（mHealth）应用程序，该程序可以测量个体的步行运动，并根据个体的健康状况分析不同个体步行运动的差异。智能手机的运动传感器用于测量步行运动，并通过计算每个 xyz 旋转的变化来分析旋转矩阵数据，这显示了随时间变化的与步行相关运动数据中的变量。
从 3 名参与者中收集数据，即 2 名健康个体（1 名女性和 1 名男性）和 1 名患有背痛的男性。患有背痛的参与者在剧烈运动中背部受伤，但他在行走方面没有任何问题。参与者将智能手机戴在腰部中间（作为重心），然后在建筑物的室内走廊中以自己的速度直线行走。他们走了大约 400 英尺。他们在直线上走了 2-3 分钟，然后返回起始位置。智能手机中的旋转向量，通过旋转矩阵计算，用于测量人体行走时的俯仰、横滚和偏航角度。重新计算每个 xyz-旋转向量数据，以找到每个参与者步行运动的变化。
与健康参与者相比，患有背痛的男性参与者的步行平衡水平较低，旋转时 xyz 轴的变化范围较宽。患有背痛的男性参与者的 xyz 轴标准差大于健康男性参与者。此外，患有背痛的参与者的左右和前后运动的组合范围最宽。健康男性参与者的步行标准偏差小于患有背痛的男性参与者和健康女性参与者，这表明健康男性参与者的步行运动平衡良好。健康女性参与者的步行运动在旋转向量的 x-y 变化中表现出左右（x 轴）和上下（y 轴）运动的对称性，表明她的步态偏差较小。
本研究表明，我们基于智能手机传感器和旋转向量的 mHealth 应用程序可以测量不同个体的步行运动变化。需要进一步的研究来测量和比较不同年龄、性别以及健康问题或疾病类型的步行运动。该应用程序可以帮助发现影响步态的疾病患者在步态方面的差异。
