This paper is focused on the development of a control system, implemented on a parallel robot designed for the lower limb rehabilitation of bedridden stroke survivors. The paper presents the RECOVER robotic system kinematics, further implemented into the control system, which is described in terms of architecture and functionality. Through a battery of experimental tests, achieved in laboratory conditions using eight healthy subjects, the feasibility and functionality of the proposed robotic system have been validated, and the overall performance of the control system has been studied. The range of motion of each targeted joint has been recorded using a commercially available external sensor system. The kinematic parameters, namely the patient’s joints velocities and accelerations have been recorded and compared to the ones obtained using the virtual model, yielding a very small difference between them, which provides a validation of the RECOVER initial design, both in terms of mechanical construction and control system.

Motor disability is a key feature of many neurological diseases, influencing the social roles of affected patients and their ability to perform daily life activities. Current rehabilitation capacities are overwhelmed by the age-related increase of motor dysfunctions seen, for example, in stroke, extrapyramidal or neuromuscular diseases. As the patient to rehabilitation personnel ration increases, robotic solutions might establish the possibility to rapidly satisfy the increasing demand for rehabilitation. This paper presents an inaugural exploratory study which investigates the interchangeability of a novel experimental robotic rehabilitation device system with classical physical therapy, using a multimodal neurophysiological assessment of the motor system—quantitative electroencephalogram (EEG), motor conduction times and turn/amplitude analysis. Preliminary results show no significant difference between the two methods; however, a significant effect of the therapy was found on different pathologies (beneficial for vascular and extrapyramidal, or limited, and only on preventing reduction of joint movements in neuromuscular).

The use of robotic systems in physical rehabilitation protocols has become increasingly attractive and has been given more focus in the last decade as a result of the high prevalence of motor deficits in the population, which is linked to an overburdened healthcare system. In accordance with current trends, three robotic devices have been designed, called ParReEx Elbow, ParReEx Wrist, and ASPIRE, which were designed to improve upper-limb medical recovery (shoulder, elbow, forearm, and wrist). The three automated systems were tested in a hospital setting with 23 patients (12 men and 11 women) suffering from motor deficits caused by various neurological diseases such as stroke, Parkinson’s disease, and amyotrophic lateral sclerosis (ALS). The patients were divided into three groups based on their pathology (vascular, extrapyramidal, and neuromuscular). Objective clinical measures, such as the Medical Research Council (MRC) scale, goniometry, and dynamometry, were used to compare pre- and post-rehabilitation assessments for both robotic-aided and manual physical rehabilitation therapy. The results of these tests showed that, with the exception of a few minor differences in muscular strength recovery, the robotic-assisted rehabilitation methods performed equally as well as the manual techniques, though only minor improvements were validated during short-term rehabilitation. The greatest achievements were obtained in the goniometric analysis where some rehabilitation amplitudes increased by over 40% in the vascular group, but the same analysis returned regressions in the neuromuscular group. The MRC scale analysis returned no significant differences, with most regressions occurring in the neuromuscular group. The dynamometric analysis mostly returned improvements, but the highest value evolution was 19.07%, which also in the vascular group. While the results were encouraging, more research is needed with a larger sample size and a longer study period in order to provide more information regarding the efficacy of both rehabilitation methods in neurological illnesses.

The paper presents the design optimization of the ASPIRE spherical parallel robot for shoulder rehabilitation following clinical evaluation and clinicians’ feedback. After the development of the robotic structure and the implementation of the control system, ASPIRE was prepared for clinical evaluation. A set of clinical trials was performed on 24 patients with different neurological disorders to obtain the patient and clinician acceptance of the rehabilitation system. During the clinical trials, the behavior of the robotic system was closely monitored and analyzed in order to improve its reliability and overall efficiency. Along with its reliability and efficiency, special attention was given to the safety characteristics during the rehabilitation task.

The research aimed to evaluate the efficacy of the NeuroAssist, a parallel robotic system comprised of three robotic modules equipped with human–robot interaction capabilities, an internal sensor system for torque monitoring, and an external sensor system for real-time patient monitoring for the motor rehabilitation of the shoulder, elbow, and wrist. The study enrolled 10 consecutive patients with right upper limb paresis caused by stroke, traumatic spinal cord disease, or multiple sclerosis admitted to the Neurology I Department of Cluj-Napoca Emergency County Hospital. The patients were evaluated clinically and electrophysiologically before (T1) and after the intervention (T2). The intervention consisted of five consecutive daily sessions of 30–45 min each of 30 passive repetitive movements performed with the robot. There were significant differences (Wilcoxon signed-rank test) between baseline and end-point clinical parameters, specifically for the Barthel Index (53.00 ± 37.72 vs. 60.50 ± 36.39, p = 0.016) and Activities of Daily Living Index (4.70 ± 3.43 vs. 5.50 ± 3.80, p = 0.038). The goniometric parameters improved: shoulder flexion (70.00 ± 56.61 vs. 80.00 ± 63.59, p = 0.026); wrist flexion/extension (34.00 ± 28.75 vs. 42.50 ± 33.7, p = 0.042)/(30.00 ± 22.97 vs. 41.00 ± 30.62, p = 0.042); ulnar deviation (23.50 ± 19.44 vs. 33.50 ± 24.15, p = 0.027); and radial deviation (17.50 ± 18.14 vs. 27.00 ± 24.85, p = 0.027). There was a difference in muscle activation of the extensor digitorum communis muscle (1.00 ± 0.94 vs. 1.40 ± 1.17, p = 0.046). The optimized and dependable NeuroAssist Robotic System improved shoulder and wrist range of motion and functional scores, regardless of the cause of the motor deficit. However, further investigations are necessary to establish its definite role in motor recovery.

Robust exercise recognition is essential for robot-assisted lower-limb rehabilitation, where misclassifications of sensor-derived movements can degrade therapy execution and supervision. This study proposes a novel hybrid weighted stacking ensemble to increase the efficiency of the intelligent module of the LegUp parallel robotic system for lower limb rehabilitation. The approach combines a Residual Multilayer Perceptron (ResMLP) and an optimized Kernel Extreme Learning Machine (KELM), where model hyperparameters are tuned using Optuna and the base-model probability outputs are fused through optimized weighting and a meta-learner. Experiments were conducted on a five-class dataset built from nine IMU orientation features acquired from three sensors placed on the healthy limb. Four meta-learners were evaluated (Logistic Regression, Random Forest, Gradient Boosting, and AdaBoost), with AdaBoost providing the best overall performance. To further assess the robustness and generalization capability of the proposed approach, a 5-fold cross-validation procedure was performed for the ResMLP, KELM, and the hybrid ensemble models. The proposed stacking hybrid ensemble consistently surpassed the performance of the strongest individual classifiers as well as the original LegUp Multilayer Perceptron model. These results indicate that combining residual learning with kernel-based classification in a weighted stacking framework yields a stable and high-performing solution for multi-class rehabilitation exercise recognition.