Pneumatic muscle actuators (PMAs), a kind of soft/compliant actuators, have been attracted a great deal of attention in the studies of rehabilitation robots. However, the nonlinearities, uncertainties, hysteresis, and time-varying features of PMAs bring a lot of difficulties in their high-precision trajectory tracking tasks. In this paper, an echo state Gaussian process-based nonlinear model predictive control (ESGP-NMPC) is designed for the PMAs. The proposed strategy is comprised of an ESGP, which is suitable for modeling unknown nonlinear systems as well as measuring their uncertainties, and a gradient descent optimization algorithm for calculating the control signal sequences. Based on the Lyapunov theorem, characteristics of the closed-loop system are analyzed to guarantee the asymptotical stability. Both simulations and physical experiments are carried out to illustrate the validity of the proposed control strategy. Compared with other conventional methods, the ESGP-NMPC can achieve a better model fitting for the PMA and control performance for the high-precision tracking tasks.

Since manual inspection of analog instruments is inefficient, many computer vision-based automatic reading systems have been proposed recently. However, most of them use fixed cameras, which are costly due to the large number of used cameras. Although some other systems adopting the pan-tiltzoom camera and the movable inspection robot can avoid using plenty of cameras, they have to overcome high computational cost in aligning the camera to the tested instrument. Meanwhile, most existing systems are instrument type dependent, and hence cannot handle multiple types of instruments simultaneously. In this paper, first, based on an inspection robot, an automatic reading system equipped with a pan-tilt-zoom camera is designed for different types of round-shape analog instruments. Then, a fast camera alignment algorithm based on visual servo is proposed, in which YOLOv3 is applied and improved to locate the instrument, and guide the camera to iteratively align to the instrument. Finally, a monocular-vision pointer reconstruction algorithm is proposed to accurately read the instrument. Experimental results demonstrated that our proposed system is fast and reliable in the camera alignment process, and is effective in reading different types of analog instruments during the robot-based inspection.

In rehabilitation practice, motivating patients with neurological injuries to actively increase muscle activity and ensure their safety are important. Therefore, this study proposed a position-constrained assist-as-needed (AAN) control method for rehabilitation robots. A human–robot interaction system with position constraints was first established based on prescribed performance. Aiming at implementing the AAN strategy, the robot assistance level metric (RALM), a constructed global continuous differentiable function incorporating dead zone and saturation characteristics, was introduced to quantify the robotic assistance and facilitate seamless operation. To bridge the gap between the position constraints and the AAN strategy, a sliding manifold was constructed for the constrained human–robot dynamic system, where RALM was regarded as a weight factor to achieve a human-dominated mode, a robot-dominated mode, and their smooth transition, regarded as a human–robot shared mode. The stability of the closed-loop system was guaranteed by using the Lyapunov theory, and the proposed controller was verified by several physical experiments on a knee exoskeleton driven by pneumatic muscles.

In Pneumatic Muscle Actuators (PMAs)-driven robotic applications, there might exist unpredictable shocks which lead to the sudden change of desired trajectories and large tracking errors. This is dangerous for physical systems. In this paper, we propose a novel adaptive proxy-based robust controller (APRC) for PMAs, which is effective in realizing a damped response and regulating the behaviors of the PMA via a virtual proxy. Moreover, the integration of the APRC and the nonlinear disturbance observer (NDO) further handles the system uncertainties/disturbances and improves the system robustness. According to the Lyapunov theorem, the tracking states of the closed-loop PMA control system are proven to be globally uniformly ultimately bounded through two motion phases. Extensive experiments are conducted to verify the superior performance of our approach, in multiple tracking scenarios.

This article presents a single-layer learning-based predictive control strategy for pneumatic muscle actuators (PMAs)-driven lower limb exoskeleton. Although PMAs are promising for rehabilitation robots, they suffer from nonlinearities, unmodeled uncertainties, hysteresis, etc. As a consequence, the mechanism actuated by PMAs rarely involves complex dynamics, and the related precise control remains a challenging problem. Hence, considering the global approximation capability of neural networks, we use an echo state network (ESN) to approximate the dynamics of the PMAs-driven exoskeleton with a nonlinear autoregressive exogenous model and forecast its behaviors by constructing training and testing data sets. Through the model predictions, the idea of single-layer learning solves a quadratic programming problem based on the principle of predictive control over a finite future horizon. After that, the control strategy turns out to be asymptotically stable when the ESN is capable of approximating the dynamics of the exoskeleton. Passive gait training experiments are conducted with six healthy subjects to verify the effectiveness of the proposed control strategy. Compared with the traditional strategies, the proposed control strategy achieves higher tracking accuracy for passive gait training tasks.