Affordable Robot Assisted Gait Rehabilitation for Stroke Survivors

Faculty Development Competitive Research Grant Program 2018-2020

The present research is aimed at developing an affordable robot assisted physiotherapy solution for stroke survivors in order to improve their treatment outcomes. Stroke is a leading cause for lower limb disabilities in Kazakhstan and rest of the world among young and the ageing population [1-5]. In the United States of America (USA) about 795,000 people fall victim to a new stroke and about 191,000 people in Japan have first or a recurrent stroke[6]. The number of new stroke patients in China is about 200 million each year [7]. Stroke mortality rate in Kazakhstan is increasing and is one of the highest in the world with 239.9 per 100,000 compared to 75.8 and 48 in Japan and USA respectively [8]. While stroke morbidity, mortality, and recurrence rates increase with age, stroke incidences are not uncommon in the younger population these days. Stroke survivors normally have sensory and motor function disorders which often cause lower limb disability and reduced ambulation. The stroke rehabilitation and hospitalization costs are one of the highest for all injuries and yet there is no universally accepted rehabilitation approach [9, 10]. Conventional manual therapies, namely, physical therapy (PT) and occupation therapy (OT) greatly depend on the therapists’ experience and with time it is becoming difficult to meet the ever increasing requirements of high-intensity and repetitive training. Body weight supported (BWS) manually assisted treadmill training is in practice for more than 20 years [11, 12] and has shown improvements in patients [13-20]. However, to train patients in standing, walking and performing balance activities, arduous efforts from a team of physiotherapists is required. Owing to the shortage of expert therapists, the treatment is prolonged and may not be effective [21]. The evaluation of present treatments suggests that only one quarter of patients in rehabilitation regain some ability to walk by the time of discharge [22, 23].To improve upon the treatment outcomes and save therapists from physical fatigue, robot assisted rehabilitation solutions have been actively researched during past two decades [24]. Robots, supervised by therapists, can deliver well controlled repetitive and prolonged training sessions with enhanced accuracy, objectivity and effectiveness with reduced human intervention [25-31]. Robot assisted rehabilitation treatment can help therapists to supervise many subjects simultaneously, analyze the recorded data in order to evaluate the patient’s health and make an objective decision for future course of treatment. Consequently, many rehabilitation robots have been proposed by researchers during last two decades for automated training [32-41]. The first modern robotic gait training orthosis LOKOMAT, now commercially available, was developed in the late 90’s [24]. Several other robotic gait training orthoses have also been developed during the last decade [42-45]. The ongoing research work at the Department of Electrical and Electronics Engineering, Nazarbayev University focuses on the rehabilitation of stroke survivors by developing an affordable robot assisted gait rehabilitation system. Our research suggests that there are at least two major issues which need further research in order to exploit the potential of robotic exoskeletons in gait rehabilitation treatment [46, 47]. Firstly, designs of the existing robot mechanisms do not remain kinematically compliant with the human limbs while in motion. This further means that the robot joints and the human anatomical joints do not remain aligned during actuation and as such there is a need to investigate into bio-inspired robot designs which have increased degrees of freedoms (dof). A design which remains kinematically compliant with human motions shall provide effective and safe gait training. The alignment of these robotic orthoses, especially, hip and ankle joint orthoses with anatomical joints, present a major design challenge because the hip and ankle are complex anatomical joints. An attempt has been made to design a knee robotic orthoses based on the anatomical joint features, but no such attempt has been reported for ankle and hip joints [48]. Various mechanisms have been proposed for upper limb orthoses that can provide better alignment with anatomical joints [49, 50]. These mechanisms can be modified and adapted to the lower limb robotic orthoses in order to provide better joint alignment. Secondly, majority of the control schemes proposed for these robotic orthoses work on the basis of trajectory tracking control (assist only) which forces the subject’s limbs on predefined trajectories without taking into account the patient’s disability level [24]. Neurologically impaired patients often suffer from severe spasms. These orthoses produce large forces in response to the uncontrolled motions produced by spasms (and resulting position errors) and as a result the patient may feel pain and discomfort [51, 52]. Since the trajectory tracking control operates in a robot in-charge mode, patients are not able to effectively contribute and as a result the recuperation is slow [53]. There is a strong need to develop and implement assist as needed (AAN) control schemes for these orthoses in order to provide customized gait training and allow subjects’ active contribution. Investigators of this project have proposed an AAN control approach whereby it was reported that the controller was able to modify robotic assistance based on the voluntary participation of human subjects [47]. Nevertheless, this approach needs to be consolidated subsequent to further investigations and validations. Moreover, this AAN control scheme was only evaluated on healthy subjects and in order to establish its therapeutic efficacy, rigorous clinical trials with neurologically impaired subjects are necessary. There are few other issues which we endeavor to investigate are, higher costs of orthoses [54] and undesired use of stiff and heavy actuators [44, 55-58] which further increases overall orthosis weight.
In order to advance the present state of robot assisted physiotherapy, we plan to develop a new design of robotic orthosis which will be more flexible (employing more degrees of freedoms), light weight and will use ‘pneumatic muscle actuators (PMA)’ which behave similar to human skeletal muscles [59, 60]. These actuators add compliance to the actuation mechanism which helps in absorbing large position errors and ensures safety besides providing impact resistance during heel strike and stepping. We also plan to implement our newly devised AAN control approach on this robotic orthosis in order to estimate the level of individual impairment and adjust the robotic assistance accordingly so that the patient can contribute more towards the rehabilitation process [29]. One such kind of training strategy, referred to as impedance control, has been implemented on LOKOMAT [61, 62] but this attempt has added an extra layer of control complexity [63] and the switching of robot assistance (low to high) is done manually.
The performance of robotic physiotherapy needs to be evaluated in real time and necessary feedback from the system be obtained in order to perform an evidence based treatment. We propose to perform dynamic simulations to identify musculoskeletal functions of subjects and also develop understanding of correlation between muscle coordination abnormalities and gait pattern deviations [64]. Dynamic simulations are able to provide knowledge of the etiology of movement disorders and more effective treatments subsequently [65-67]. Apart from this, EMG measurements of human subjects [68, 69] shall also be recorded during gait training experiments [24]. Dynamic simulations can help in overcoming the limitations of EMG measurements in studying muscle activations patterns [65, 66, 70, 71].
Impacts of the results: The impacts of this project and related research shall be manifold in the arena of medical devices, stroke treatment and in enhancing extant knowledge of human walking. Specifically in Kazakhstan where the population density is low and people live in distant places, the robotic devices can play an important role in helping small community of therapists to connect to the patients virtually and serve them remotely. The proposed robotic orthosis also has potential for commercialization and one can envision some financial opportunity for the country. The research shall also provide platform for the young scientists in the country to learn the process of design and development of medical robots which is vital in the pursuit of making them future entrepreneurs.
Novelty in the proposed research:
The proposed research eventually results in a Gait rehabilitation orthosis which will be used in clinics for the treatment of stroke survivors. This new design will be cost effective and will also be technologically advanced. The proposed orthosis shall employ ‘pneumatic muscle actuators (PMA)’ which behave similar to the skeletal muscles and therefore the actuation of this design will be soft and compliant (without jerks). The proposed design will have enhanced degrees of freedom which will make it more flexible than existing robots. We will also implement ‘assist-as-needed’ control architecture for providing seamless adaptive robotic assistance during gait training. The PI and his collaborators have already done some preliminary work with the proposed controller [47]. This controller will be able to adapt the robotic assistance to the patient’s disability levels. Extending maximum assistance to severely impaired subjects and minimum assistance for subjects with less severe impairments can enhance patient involvement and positively contribute towards earlier recovery from the disabilities.

So far the literature review has been completed and a journal article submitted in the reputed Journal of Engineering in Medicine (SAGE publisher) [1, 2].
The kinematic and dynamic analysis of the robot mechanism is complete and we have a started to work on the controller design.
A new algorithm for the design optimization of gait robot has recently been accepted for publication in IEEE ACCESS [3].