Deep-Learning-Based Emergency Stop Prediction for Robotic Lower-Limb Rehabilitation Training Systems

Intervention Effect of Rehabilitation Robotic Bed Under Machine Learning Combined With Intensive Motor Training on Stroke Patients With Hemiplegia

It was aimed to discuss the effect of bed-type rehabilitation robots under machine learning combined with intensive motor training on the motor function of lower limbs of stroke patients with hemiplegia. A total of 80 patients with stroke hemiplegia were taken as the subjects, who all had a course of treatment for less than 6 months in the Rehabilitation Medicine Department of Ganzhou Hospital. These patients were divided into the experimental group (40 cases) and the control group (40 cases) by random number method. For patients in the control group, conventional intensive motor training was adopted, whereas the conventional intensive motor training combined with the bed-type rehabilitation robot under machine learning was applied for patients in the experimental group. Fugl-Meyer Assessment of Lower Extremity (FMA-LE), Rivermead Mobility Index (RMI), and Modified Barthel Index (MBI) were used to evaluate the motor function and mobility of patients. The human–machine collaboration experiment system was constructed, and the software and hardware of the control system were designed. Then, the experimental platform for lower limb rehabilitation training robots was built, and the rehabilitation training methods for stroke patients with hemiplegia were determined by completing the contact force experiment. The results showed that the prediction effect of back-propagation neural network (BPNN) was better than that of the radial basis neural network (RBNN). The bed-type rehabilitation robot under machine learning combined with intensive motor training could significantly improve the motor function and mobility of the lower limbs of stroke patients with hemiplegia.