Jiaxu Wang

The lower limb exoskeleton can improve mobility and safety during rehabilitation for human. However, most current exoskeleton systems are not capable of providing variable joint stiffness in response to changing external demands. In this paper, a knee exoskeleton based on the series elastic actuator (SEA) is designed for safe human-computer interaction. The structural dimensions of the exoskeleton actuation mechanism were optimized based on gait biomechanics to ensure stability and compactness. While maintaining the mechanism range of motion (ROM), this optimization ensures that less peak force is required during the gait cycle. However, the insertion of series elastic actuators inevitably brings new challenges for high precision control of the exoskeleton, such as the problems of modeling errors, compliance, friction, and external disturbances in the exoskeleton joint. To achieve high precision control of the exoskeleton, an extended disturbance observer (EDO) based command filtered backstepping control (CFBC) of the knee exoskeleton is developed. The effective observation of friction, external disturbances, and modeling errors in the system is obtained by the EDO. Compared with conventional backstepping control, the CFBC can not only solve the “explosion of complexity” problem through a command filter but also reduce filter errors by an error compensation mechanism. Based on the Lyapunov stability, all signals in the closed-loop system are semiglobal uniformly ultimately bounded. Finally, comparison simulation results demonstrate the effectiveness of the proposed control approach.

Transmission compliance and hysteresis behaviors are associated with harmonic drives. The experimental observation of a harmonic drive showed that these behaviors are significantly degraded with increasing service time, and their accurate modeling is expected to improve the performance of the control system of harmonic drive-based devices such as robot manipulators. In this paper, a new model is proposed to capture transmission compliance and hysteresis and their degradation. The phenomenon is represented through a combination of the nonlinear stiffness component and the micro-sliding friction in the tooth engagement area. The proposed model considered the multi-tooth meshing and the interference effects of the reducer. The parameters of the model were identified using optimization techniques. Numerical simulations and experimental data using a specialized harmonic drive test device were compared to demonstrate the proposed model.

In this study, a physics-based fatigue wear model is proposed to evaluate the reliability and to predict the life of cumulative micropitting wear for lubricated conformal contacts on rough surfaces. The surface normal load, mean film thickness, and frictional shear traction are simulated by a mixed elastohydrodynamic lubrication (EHL) model for a stress prediction model to calculate the average maximum Hertzian pressure of contact asperities and unit with the statistical contact model and dynamic contact model to obtain the asperity stress cycle number. The wear formula is established through combining a micropitting life prediction model of surface asperities and a mean micropitting damage constant of asperities. The four dominant aspects affecting wear behaviors of the surface contact pairs, working conditions, structure and surface topographies, material properties and lubrication conditions are all taken into account in the model. It is a high-fidelity and comprehensive model that can be used to analyze and optimize the tribological design of rolling–sliding pairs in machinery. The micropitting fatigue wear modeling scheme is validated by comparison of theoretical calculations and available experimental wear data.