Volume 70, Issue 11 - Special Issue: AUTOMED 2021: Automation in Medical Technology; Guest Editors: Georg Rauter, Thomas Seel and Philipp Rostalski

Relative comparison of clinical scores to measure the effectiveness of neuro-rehabilitation therapy is possible through a series of discrete measurements during the rehabilitation period within specifically designed task environments. Robots allow quantitative, continuous measurement of data. Resulting robotic scores are also only comparable within similar context, e.g. type of task. We propose a method to decouple these scores from their respective context through functional orthogonalization and compensation of the compounding factors based on a data-driven sensitivity analysis of the user performance. The method was validated for the established accuracy score with variable arm weight support, provoked muscle fatigue and different task directions on 6 participants of our arm exoskeleton group on the ANYexo robot. In the best case, the standard deviation of the assessed score in changing context could be reduced by a factor of 3.2. Therewith, we paved the way to context-independent, quantitative online assessments, recorded autonomously with robots.

In this paper, a testing for highly automated function (HAF) is adapted from the automotive industry to therapeutic medical devices. It contains different steps to achieve a safety argumentation: First, scenarios of interest (SoI) based on a systematic generalization of failure mode and effect analysis (FMEA) are identified, then the concrete scenarios are generated using design of experiment (DoE). These scenarios are simulated virtually and physically and are then evaluated. The procedure is explained with the use of examples.

This article presents an approach for model-based and personalized determination of inspiratory pressure and inspiratory time in pressure-controlled ventilation. Lung mechanics are strongly dependent on weight and affected by lung diseases such as COPD. Based on a simplified model of the lung with weight- and COPD-specific parameter selection and consideration of ventilation guidelines, the appropriate inspiratory pressure values are calculated. The results illustrate the effect of BMI on the computed optimal pressure-volume ratio. The findings provide a first step towards individualized decision support systems taking into account additional effects like BMI and specific lung diseases.

Roller pumps are used in medicine to pump blood in extracorporeal blood circuits. Blood is subjected to mechanical stress in these peristaltic pumps, due to the complete tube occlusion. This stress causes damage to red blood cells (hemolysis). We present a novel peristaltic pumping principle that addresses these issues. In contrast to conventional roller pumps, the tube is stimulated by a circulating eccentric oscillation. Therefore, this type of pump does not require complete tube occlusion. The blood is less mechanically stressed and hemolysis is reduced by 55% compared to a roller pump.

Bone screws are used in many orthopaedic procedures. For low-density cancellous and osteoporotic bone, the torque margin for bone screw insertion can be low; over-tightening will strip the threads, while under-tightening may allow screw loosening over time. A model-based system of automatic individualised torque regulation has been proposed, however current efforts assume speed-independence. This publication addresses the lack of knowledge on the speed dependance of insertion torque. Therefore, a bone screw was inserted at various speeds between 2.5 and 25 rpm into 4 different densities of polyurethane foam synthetic bone. A statistically significant upwards trend was found for all materials ( p < 0.05). These findings can be used to correct identified bone strength values from previous models, although investigating the physical cause may improve correction schemes.

Laser photocoagulation is a technique applied in the treatment of retinal disease, which is often done manually or using simple control schemes. We pursue an optimization-based approach, namely Model Predictive Control (MPC), to enforce bounds on the peak temperature and, thus, to ensure safety during the medical treatment procedure – despite the spot-dependent absorption of the tissue. The desired laser repetition rate of 1 kHz is renders the requirements on the computation time of the MPC feedback a major challenge. We present a tailored MPC scheme using parametric model reduction, an extended Kalman filter for the parameter and state estimation, and suitably tuned stage costs and verify its applicability both in simulation and experiments with porcine eyes. Moreover, we give some insight on the implementation specifically tailored for fast numerical computations.

Frequent rehabilitation exercises can accelerate the recovery of patient’s hand impairments after stroke. In conjunction with conventional therapy, the use of robot-assisted training has been proposed to increase the frequency of the latter, thus improving the overall recovery of patients. However, the COVID-19 pandemic has reduced or even halted related programs in clinical rehabilitation centers that often utilize costly, complex, and non-transportable robotic devices. A possible therapy alternative is using low-cost in-home solutions which integrate well in home-based settings due to reduced size, weight, and complexity. Therefore, we propose a new portable hand training and assessment device which supports hand opening/closing and wrist exercises. The device also enables tasks related to the identification of vibration stimuli to be used both for training as well as for assessment of sensory-motor hand function of patients. To this end, a cable-driven capstan transmission mechanism, a controller to regulate the extension angle, and touch pads equipped with vibration motors were designed and integrated into a hand-held device. Initial user studies indicate that the prototype can support stroke patients in extending their fingers. In further experiments targeting the identification of vibration stimuli, assessment results comparable with those obtained via a standard clinical somatosensory assessment test were achieved. Given these initial findings, our low-cost system shows the potential to allow in-home rehabilitation exercises in daily life, thus maximizing exposure and frequency, even during a forced lockdown caused by a pandemic.

For many elderly, wheeled rollators play an essential role in everyday life’s mobility. In addition, exoskeletons are increasingly used in daily mobility as well as in rehabilitation therapy. We present a motorized rollator to combine the advantages of both support devices. Based on the developed hardware, we design two model-based control concepts. The first relies on distance measurement with ultrasonic sensors, which enables contactless navigation of the rollator, to allow balance training with the exoskeleton. The second control concept enables the balancing of the rollator on its rear wheels as a first step to overcome stairs. Through experiments on the real system, we demonstrate and discuss the functionality and reliability of both control concepts.

In this dissertation, a novel class of model structures and associated training algorithms for building data-driven nonlinear state space models is developed. The new identification procedure with the resulting model is called local model state space network (LMSSN). Furthermore, recurrent neural networks (RNNs) and their similarities to nonlinear state space models are elaborated on. The overall outstanding performance of the LMSSN is demonstrated on various applications.