Optimum human input: Exoskeletons can be used to augment human abilities—for example, to lift very heavy loads or to provide greater endurance. For each user, though, a device will need to be adjusted for optimum effect, which can be time-consuming. Zhang et al. show that the human can be included in the optimization process, with real-time adaptation of an ankle exoskeleton (see the Perspective by Malcolm et al. ). By using indirect calorimetry to measure metabolic rates, the authors were able to adjust the torque provided by the device while users were walking, running, and carrying a load. Science , this issue p. 1280 ; see also p. 1230

Lower-limb exoskeletons capable of comfortably applying high torques at high bandwidth can be used to probe the human neuromuscular system and assist gait. We designed and built two tethered ankle exoskeletons with strong lightweight frames, comfortable three-point contact with the leg, and series elastic elements for improved torque control. Both devices have low mass (<; 0.88 kg), are modular, structurally compliant in selected directions, and instrumented to measure joint angle and torque. The exoskeletons are actuated by an off-board motor, and torque is controlled using a combination of proportional feedback and damping injection with iterative learning during walking tests. We tested closed-loop torque control by commanding 50 N·m and 20 N·m linear chirps in desired torque while the exoskeletons were worn by human users, and measured bandwidths greater than 16 Hz and 21 Hz, respectively. During walking trials, we demonstrated 120 N·m peak torque and 2.0 N·m RMS torque tracking error. These performance measures compare favorably with existing devices and with human ankle musculature, and show that these exoskeletons can be used to rapidly explore a wide range of control techniques and robotic assistance paradigms as elements of versatile, high-performance testbeds. Our results also provide insights into desirable properties of lower-limb exoskeleton hardware, which we expect to inform future designs.

Few comparisons have been performed across torque controllers for exoskeletons, and differences among devices have made interpretation difficult. In this study, we designed, developed and compared the torque-tracking performance of nine control methods, including variations on classical feedback control, model-based control, adaptive control and iterative learning. Each was tested with four high-level controllers that determined desired torque based on time, joint angle, a neuromuscular model, or electromyography. Controllers were implemented on a tethered ankle exoskeleton with series elastic actuation. Measurements were taken while one human subject walked on a treadmill at 1.25 m·s ^-1 for one hundred steady-state steps. The combination of proportional control with damping injection and iterative learning resulted in the lowest errors for all high-level controllers. With time-based desired torque, root-mean-squared errors were 0.6 N·m (1.3% of peak desired torque) step by step, and 0.1 N·m (0.2%) on average. These results indicate that model-free, integration-free feedback control is suited to the uncertain dynamics of the human-robot system, while iterative learning is effective in the cyclic task of walking.

Each year, stroke and traumatic brain injury leave millions of survivors with motion control loss, which results in great demand for recovery training. The great labor intensity in traditional human-based therapies has recently boosted the research on rehabilitation robotics. Existing controllers for rehabilitative robotics cannot solve the closed-loop system stability with uncertain nonlinear dynamics and conflicting human-robot interactions. This paper presents a theoretical framework that establishes the passivity of the closed-loop upper-limb rehabilitative robotic systems and allows rigorous stability analysis of human-robot interaction. Position-dependent stiffness and position-dependent desired trajectory are employed to resolve the possible conflicts in motions between patient and robot. The proposed method also realizes the “assist-as-needed” strategy. In addition, it handles human-robot interactions in such a way that correct movements are encouraged and incorrect ones are suppressed to make the training process more effective. While guaranteeing these properties, the proposed controller allows parameter adjustment to provide flexibility for therapists to adjust and fine tune depending on the conditions of the patients and the progress of their recovery. Simulation and experiment results are presented to illustrate the performance of the method.

Using “human-in-the-loop” (HIL) optimization can obtain suitable exoskeleton assistance patterns to improve walking economy. However, there are differences in these patterns under different gait conditions, and currently most HIL optimizations use metabolic cost, which requires long periods to be estimated for each control law, as the physiological objective to minimize. We aimed to construct a muscle-activity-based cost function and to find the appropriate initial assistance patterns in HIL optimization of multi-gait ankle exoskeleton assistance. One healthy subject walked assisted by an ankle exoskeleton under nine gait conditions and each condition was the combination of different walking speeds, ground slopes and load weights. Ten assistance patterns were provided for the subject under each gait condition. Then we constructed a cost function based on surface electromyography signals of four lower leg muscles and select the muscle weight combination by using particle swarm optimization algorithm to compose the cost function with maximum differences between different assistance patterns. The mean weights of medial gastrocnemius, lateral gastrocnemius, soleus and tibialis anterior activity under all gait conditions are 0.153, 0.104, 0.953 and 0.145, respectively. Then we verified the effectiveness of this cost function by optimization and validation experiments conducted on four subjects. Our results are expected to guide the selection of muscle-activity-based cost functions and improve the time efficiency of HIL optimization.