RESUMEN
Individuals with acute neurological or limb-related disorders may be temporarily bedridden and unable to go to the physical therapy departments. The rehabilitation training of these patients in the ward can only be performed manually by therapists because the space in inpatient wards is limited. This paper proposes a bedside cable-driven lower-limb rehabilitation robot based on the sling exercise therapy theory. The robot can actively drive the hip and knee motions at the bedside using flexible cables linking the knee and ankle joints. A human-cable coupling controller was designed to improve the stability of the human-machine coupling system. The controller dynamically adjusts the impedance coefficient of the cable driving force based on the impedance identification of the human lower-limb joints, thus realizing the stable motion of the human body. The experiments with five participants showed that the cable-driven rehabilitation robot effectively improved the maximum flexion of the hip and knee joints, reaching 85° and 90°, respectively. The mean annulus width of the knee joint trajectory was reduced by 63.84%, and the mean oscillation of the ankle joint was decreased by 56.47%, which demonstrated that human joint impedance identification for cable-driven control can effectively stabilize the motion of the human-cable coupling system.
RESUMEN
Introduction: With the aggravation of aging and the growing number of stroke patients suffering from hemiplegia in China, rehabilitation robots have become an integral part of rehabilitation training. However, traditional rehabilitation robots cannot modify the training parameters adaptively to match the upper limbs' rehabilitation status automatically and apply them in rehabilitation training effectively, which will improve the efficacy of rehabilitation training. Methods: In this study, a two-degree-of-freedom flexible drive joint rehabilitation robot platform was built. The forgetting factor recursive least squares method (FFRLS) was utilized to estimate the impedance parameters of human upper limb end. A reward function was established to select the optimal stiffness parameters of the rehabilitation robot. Results: The results confirmed the effectiveness of the adaptive impedance control strategy. The findings of the adaptive impedance control studies showed that the adaptive impedance control had a significantly greater reward than the constant impedance control, which was in line with the simulation results of the variable impedance control. Moreover, it was observed that the levels of robot assistance could be suitably modified based on the subject's different participation. Discussion: The results facilitated stroke patients' upper limb rehabilitation by enabling the rehabilitation robot to adaptively change the impedance parameters according to the functional status of the affected limb. In clinic therapy, the proposed control strategy may help to adjust the reward function for different patients to improve the rehabilitation efficacy eventually.