6 May 2024 in IEEE Sensors Journal
With the wide application of lower limb exoskeleton robots in various fields, problems such as individual motion feature recognition have gradually come to the fore. Therefore, this paper carries out a study on motion pose acquisition and action recognition based on IMU sensors. By using a motion capture system to analyze the motion movements of the human lower limbs and establishing a human lower limb motion dataset, the motion laws of the lower limb joints were extracted. We also used IMU sensors for real-time motion data acquisition. The data obtained from the two acquisition methods were compared to ensure the reliability of the IMU sensor data acquisition. To improve the accuracy of data acquisition, the IMU sensor data were processed using Kalman filtering. Based on the DTW algorithm, we carry out simulation experiments to analyze the motion data collected by IMU in real-time compared with the motion dataset, to achieve the real-time recognition of human motion movements. This research can solve the problem of individual motion feature recognition in the application of lower limb exoskeleton robots and provide strong support for research and application in rehabilitation and assisted walking.
6 April 2020 in IEEE Access
For the better design and monitoring of gearboxes operating under nonstationary conditions, the dynamic interaction behavior of an electric motor drive multistage gear set is investigated in this study. Initially, dynamic mathematical models of an asynchronous motor and a multistage gear set are established by using the equivalent circuit method and the lumped parameter method, respectively. An electromechanical model coupling the asynchronous motor and the gearbox is subsequently developed, which is suitable for nonstationary conditions. The system transient responses under an impact external load are analyzed, and the mechanical torsional-vibration and electric current response are compared. Finally, the dynamic load propagation mechanism and the interaction effect on different gears are examined. The results demonstrate that the vibration frequencies of the multistage gearbox in steady and transient states both can be captured in the current. Electric current can be applied to reflect the external load variation and the internal vibration status of the driven gearbox under stationary or nonstationary conditions. The mutual excitation effects between adjacent gear pairs change the respective dynamic characteristics of the gears, and gears with small nominal loads are susceptible.