A Novel Soft Robotic Exoskeleton System for Hand Rehabilitation and Assistance Purposes

Abstract

During the last decade, soft robotic systems, such as actuators and grippers, have been employed in various commercial applications. Due to the need to integrate robotic mechanisms into devices operating alongside humans, soft robotic systems concentrate increased scientific interest in tasks with intense human–robot interaction, especially for human-exoskeleton applications. Human exoskeletons are usually utilized for assistance and rehabilitation of patients with mobility disabilities and neurological disorders. Towards this direction, a fully functional soft robotic hand exoskeleton system was designed and developed, utilizing innovative air-pressurized soft actuators fabricated via additive manufacturing technologies. The CE-certified system consists of a control glove that copies the motion from the healthy hand and passes the fingers configuration to the exoskeleton applied on the affected hand, which consists of a soft exoskeleton glove (SEG) controlled with the assistance of one-axis flex sensors, micro-valves, and a proportional integral derivative (PID) controller. Each finger of the SEG moves independently due to the finger-dedicated motion control system. Furthermore, the real-time monitoring and control of the fabricated SEG are conducted via the developed software. In addition, the efficiency of the exoskeleton system was investigated through an experimental validation procedure with the involvement of healthy participants (control group) and patients, which evaluated the efficiency of the system, including safety, ergonomics, and comfort in its usage.

1. Introduction

Nowadays, rapid development in sectors of advanced robotic systems, automation, and smart manufacturing is observed due to the 4th industrial revolution [1,2]. Heavy industries have already embedded these systems in their production in order to optimize their workflows [3]. Robotic systems and automation are currently employed in a wide spectrum of applications, either for industrial or commercial use. On top of this, additive manufacturing technologies allow the customization of the produced products depending on the customer’s needs [4,5]. Hence, the combination of these systems offers a great opportunity to utilize them in customized applications, especially in the health sector, as the needs differ from patient to patient. More specifically, in the last decade, a series of robotic systems and automation were designed and developed to reinforce the available tools in the health sector, namely robotic exoskeletons [6], drug distribution [7], health monitoring devices [8], etc. In addition, the advances in additive manufacturing technologies and feedstock materials have enabled the customized fabrication of the aforementioned products with patient-adapted designs [9] and human-friendly materials [10]. Thus, the current research presents the development and production of a novel soft robotic exoskeleton system for hand rehabilitation and assistance purposes constructed with the aid of additive manufacturing methods. Prior to providing any further details, it is essential to introduce the hand exoskeleton’s devices, classifying them and listing their advantages and disadvantages. Hand exoskeletons are divided into three categories soft exoskeletons, rigid exoskeletons, and soft-rigid exoskeletons [11]. Soft exoskeletons are safer and provide natural movement and comfortability; however, it is difficult to produce sufficient force [12]. On the other hand, rigid exoskeletons produce enough force with abnormal movement and bulk designs [13]. Finally, a recent study has proposed the development of a hybrid soft-rigid hand exoskeleton to increase the efficiency of the device. Nevertheless, some disadvantages still occur, such as the difficult application on the human hand and the lack of comfortability [14].

Many people around the world are facing mobility difficulties and neurological disorders in different parts of their bodies, which require either special care to rehabilitate or intense assistance in order to obtain partial or full functionality of their affected limb for the accomplishment of activities of daily living (ADLs) [15]. One of the most common and crucial mobility disorders concerns the partial or full disability of hand and finger movement resulting from a stroke, neurological condition, or injury [16]. More than 50 million elderly people worldwide face these kinds of motion difficulties, increasing even more if younger people with injuries or neurological conditions are taken into account [15]. Thus, there is a necessity to develop tools and devices that could facilitate the rehabilitation process or even restore to some extent or completely hand and finger function with the assistance of exoskeletons. It is essential to clarify that besides the rehabilitation use, exoskeletons could also be used in ADLs in order to assist and perform the desired range of movements by patients. Towards this direction, in the last decade, several studies have been published analyzing the kinematics of the human hand and developing advanced soft exoskeleton gloves (SEGs) [6,17,18]. One of the first and most comprehensive research on SEG’s development has been performed by Polygerinos et al. [6]. In this work, the kinematics of a human hand was observed and analyzed via electromagnetic (EM) tracking sensors. Then, a soft pneumatic actuator from elastomeric material reinforced with woven material was designed and fabricated. Afterward, five of these actuators were assembled on a fingerless glove. These actuators were connected with a portable hydraulic system (pump & controller), resulting in a portable device designed to facilitate the movement of the human hand and fingers. Furthermore, Gerez et al. [17] developed a similar device with the addition of abduction capabilities and a telescopic thumb combined with a mobile application for the control to facilitate the interaction between the user and the device. Moreover, Chu and Patterson [18] published a narrative review about soft robotic devices for hand rehabilitation and assistance tasks. In this review, more than 60 studies were examined and presented from the last decade, pointing out the plethora of soft pneumatic robotic devices with SEG configuration providing mainly flexion motion capabilities. In addition, the study highlighted the lack of integrated hand rehabilitation devices and the absence of results from clinical trials.

To this end, the objective of the current research is to present the workflow of the design, development, and examination of a fully functional and complete soft robotic exoskeleton system for hand rehabilitation and assistance purposes, which is already utilized in ongoing clinical trials in hospitals and rehabilitation centers applied to participants with hand movement difficulties. The novelty of the developed SEG focuses on the ability to monitor and control in real-time the position of each finger and also the capacity to regulate the device’s motion through the software or through a control glove that captures the desired finger motion from the healthy hand and repeated in the hand with the mobility difficulty. In this context, the function and the control of the SEG’s motion become easier, more accurate, and more realistic. Furthermore, the presented SEG reveals an innovative manufacturing procedure to rapidly fabricate customized devices for healthy applications via the utilization of additive manufacturing techniques (rapid tooling, direct manufacturing, etc.) [19,20]. To summarize, the current article presents an integrated, fully functional SEG device with real-time control and monitoring implementing modular design for the SEG via the additively manufactured parts. We present a unique system composed of an SEG that co-operates with the data glove in order to transfer the motion from the healthy hand to a hand with particular motion constraints. The design is unique, allowing bidirectional finger motions, the control is offered in real-time based on an axis flexible sensor, and the overall system is designed to respect the safety constraints imposed by the targeted application domain.

In particular, the current paper consists of five sub-sections, of which Section 2.1 lists the requirements coupled with the technical specifications of an SEG device. In Section 2.2, the employed actuator is presented with its motion mechanism and the measured experimental capabilities. Furthermore, the strategies of the developed software for this device are described in Section 2.3. In Section 3.1, the final design and assembly of the developed SEG are depicted with both the physical device and the software application. Finally, Section 3.2 summarizes the most important results from measurements of the device function coupled with the certifications and the extensive testing that has been conducted by healthy participants (control group) and patients. Figure 1 portrays the flowchart of the SEG’s device development coupled with a detailed description of each followed procedure.

2. Development of the SEG Device

2.1. Requirments and Technical Specifications

Before setting the requirements for an SEG device, it is essential to track down the clinical conditions that require the use of SEG devices. As aforementioned, this device facilitates rehabilitation after a medical condition (injury, stroke, etc.). Regardless of the medical issues that led to the disability, the cases of monoplegia and hemiplegia (Figure 2a) are the medical conditions in which SEG devices could be useful in terms of rehabilitation and assistance [21]. It is worth mentioning that according to the literature, these cases contribute a percentage of around 40% of total disability cases [21]. The main idea for the functionality of SEG is the widespread rehabilitation method of mirror therapy (MT). In detail, MT utilizes specific configurations with mirrors in order to create the illusion of the movement of the affected limb; with this technique, the sections of the brain that are related to the motion of the affected limb are activated, and the rehabilitation process is accelerated [22,23]. In traditional MT, the whole process of therapy is performed physically with the healthy limb; this offers only optical stimulus for the patient’s brain. On the other hand, the SEG device proposed in this study enables advanced MT. In detail, the utilization of the control glove on the healthy hand combined with the SEG device on the affected hand provides both optical and physical stimulus for the patient’s brain resulting in the activation of more brain sections and avoiding muscle ankyloses due to the physical movement of the affected limb [24,25]. For these reasons, the proposed SEG device appears to accelerate even more the rehabilitation process. Figure 2b depicts an indicative image of traditional MT configuration [26].

Having analyzed the medical conditions and the working principles of an SEG device, the next step is to set the main requirements and technical specifications for both the software and the hardware configuration. The first design requirement for the SEG is to be easy to use with regard to the application and its operation. For this reason, the device should be compatible with the majority of human hands, according to anthropometrics data [27]. Furthermore, both the software and the hardware should be user-friendly and harmless for the patient and healthcare professionals. Thus, the hardware was made with human-friendly and soft materials, and all electrical parts (wires, Printed Circuit Board/ PCBs, etc.) were covered with plastic cases. It is worth mentioning that in order to improve the user-friendly aspect and the rehabilitation efficiency of the device, the control of SEG could be performed via two different methods. The first method is to control the SEG through the developed software platform and the second method requires a second control glove, which is applied to the healthy hand. Then, the SEG imitates the motion of the healthy hand, stimulating the patient’s brain and offering a more realistic and physical movement for the affected limb. Moreover, the graphical user interface (GUI) of the software was designed with simple and clear graphics for each step of the procedure. One of the most important aspects of the SEG is its kinematics; the developed actuators of the SEG were designed to imitate the motion of a human finger in terms of angular range, motion speed, and physiology. In addition, in the designed SEG, the applied force of the actuators could be controlled through the working median pressure (i.e., pressurized air) in order to meet the patient’s needs. Other requirements for the developed SEG are to be lightweight (<5 kg) and easy to carry for patients and healthcare professionals. Finally, the software should be able to monitor in real-time the rehabilitation process and build libraries with the data and results of each session in order to check the progress of the patient with respect to the international legislation for personal data.

2.2. Description of the Employed Soft Actuator

One of the most crucial elements for SEGs is the utilization of soft actuators due to the fact that they perform the necessary movement in the SEG configuration. For the purposes of this work, a previously developed soft robotic actuator was utilized based on research from a previous study [28]. In that research, the processes of design, development, and fabrication were presented in-depth. More specifically, the employed actuator is a soft actuator fabricated with silicone rubber compound and reinforced with Kevlar fiber. Moreover, the kinematics of this actuator was designed to imitate the locomotion of earthworms, such as Lumbricus Terrestris [29,30,31]. Thus, it has a double half-tubular structure configuration, which enables a potential angular range from −180 to 180 degrees (extension to flexion). Furthermore, the overall structure of the actuator was developed to be comparable to an average human finger in terms of dimensions and offer multiple degrees of freedom (DoF) in order to reproduce the physical movement of a human finger (i.e., two DoF for the thumb and three DoF for the rest). However, the comprehensive advantage of this actuator is the modular 3D-printed external cover. To elaborate, the soft actuators were assembled with 3D-printed rings constructed by rigid polymer (i.e., PA12), enhancing the structural integrity of the object and allowing the free extension/flexion motion of the actuator. Moreover, the existence of the rings could offer physical constraints to avoid the undesired angles depending on the rings’ geometry and reinforce the safety of the SEG. Finally, the last advantage of the proposed actuator is the integration of a one-axis soft flex sensor inside the actuator’s structure, providing real-time data for the motion and position of the actuator. The motion of the actuator is performed through the pressured air inside the half-tubular structures, with each structure inflated for each motion direction; for example, for bending movement, the upper half-tubular structure is inflated. Therefore, the actuator’s motion could be fully controllable with an electro-pneumatic control system. Figure 3a,b show the 3D design and physical model of the actuator’s silicone rubber double half-tubular core. Also, Figure 3c portrays the configuration of the developed actuator for the SEG application.

2.3. Development of the SEG Device’s Hardware

Besides the soft actuators, there are some other components that complete the hardware of the developed SEG device, namely the control glove, the electro-pneumatic control system, the rest of the electronics, and the components of the housing. In the current section, the structures of these aforementioned components are presented thoroughly.

The control glove is an additional glove applied to the healthy hand, and its objective is to capture motion data in order to use it for the control/motion of the SEG. In the control glove, there are five one-axis soft flex sensors, namely the Bend Labs Digital Flex Sensor, one for each finger, in order to capture the finger’s motion. These bend sensors utilized the variable capacitance technology offering high accuracy and reliability with low power consumption. The other parts of the control glove are the controller configuration (PCB) and the battery. As is depicted in Figure 4, the PCB consists of the microcontroller (MCU) embedded with a WiFi module, the power management module, the sensors readout circuitry, other buttons, LEDs, and terminals. It is important to highlight that as MCU hardware, the module Esspresif ESP32-WROOM-32 was utilized. In addition, the control glove was designed to operate autonomously in terms of power due to the existence of a Lithium Polymer (LiPo) battery. Figure 4 shows the overall configuration of the control glove and its components.

The next critical module of the SEG device is the electro-pneumatic system that controls the motion of the applied soft actuators. The developed electro-pneumatic system was designed to receive pressurized air above 300 kPa and regulated at 200 kPa, which is considered the optimum value for the actuators in terms of force and durability. The pressurized air could be provided by a preinstalled pneumatic system or by a portable air pump, and the regulation of the air pressure inside the SEG device was performed with a pressure regulator. Having achieved the optimum air pressure inside the device, the next step was the selection and set-up of the electromagnetic valve system that controls each actuator. For this purpose, the TX miniature switching solenoid valve was utilized due to its small size allowing the portability of the pneumatic system. More specifically, for each actuator, three micro-valves were employed in order to allow the intake of pressured air and the outtake of the free flow in a controllable manner. Figure 5a portrays an illustration of the flexion (left side) and extension (right side) of a single actuator coupled with the schematic diagram of the pneumatic circuit. The C1 and C0 point out the condition of the valve for active or disabled status, respectively. Furthermore, the overall 15 valves are controlled via an MCU (Esspresif ESP32-WROOM-32), and the control system and the valves are positioned on a PCB. In Figure 5b, the overall pneumatic circuit is presented, coupled with the configuration of the PCB. The other electronics include power buttons, reset buttons, power and USB connector, etc. All electronics and the electro-pneumatic system were positioned on a metal case fabricated via sheet metal processing apart from the PCB and battery for the control glove that was placed into a customized 3D-printed case on the glove.

2.4. Development of the SEG Device’s Software

After the setup of the device’s hardware, the next step is the development of the corresponding software for the proper function of the SEG device. The software is divided into two pieces; the control software that retrieves the control glove data and regulates the movement of the actuator, and the graphical user interface that allows communication between the SEG device and the patient or the healthcare professional. Then, these two pieces of software were integrated into one single software platform that operates the device via the user’s commands.

The control software consists of two parts: the applied software of the control glove and the control software employed for the electro-pneumatic system that regulates the motion of the actuators. The software of the control glove is relatively simple and is executed on the MCU of the glove’s PCB. The main functionality is the detection of the finger’s angular motion, the evaluation of the finger’s angle, and the transmission of acquired data on the SEG device via the network. Figure 6a presents the main flowchart of the control glove’s software. More specifically, the user activates the control glove when a measurement is required. The bend sensors capture the movement of the finger and detect the final angle of the motion (Stable Position-SP). When the SP is achieved, the control glove transmits the measurement to the main SEG device. It is noteworthy that the process could be performed repeatedly, and all measurements are saved on the main SEG device storage for potential use.

On the contrary, the control software of the SEG is more complex and sophisticated. When the SEG device is activated, its main components are checked, i.e., bend sensors, electromagnetic valves, network connection, etc.; then, the software scans for any received data from the control glove. If there are motion data, the control sequence is executed with the corresponding states for the valves, and the desired movement is achieved with the assistance of bend sensor measurements. Finally, the software informs the user of the outcome of the movement (success or failure). Figure 6b illustrates the flowchart of the SEG device control software. Furthermore, it is worth noting that the control of the electromagnetic valves is performed with a proportional integral derivative controller (PID). The PID controller’s output is the percentage of time that the valve is activated for a predefined time period due to the valve’s binary state, i.e., on or off. Thus, the controller regulates the valve’s state depending on the desired goal angle. Moreover, the PID controller receives constant control feedback of the current angle from the bend sensors; hence it compares the current value and the goal value, and a difference value (error) is extracted. The controller minimizes that error based on the PID function protocol and the applied function coefficients [32,33,34]. The coefficients have been regulated with the empirical gain tuning achieving accurate and smooth motion without offset and overshoot errors. Figure 6c shows the experimental layout of the control and function tests (left side) coupled with the response diagram of control configuration over time (right side).

Finally, the last part of the SEG’s software is the graphical user interface (GUI) platform that facilitates the communication between the control glove and SEG with the user. The GUI application was developed in order to be able to function on commercial PCs and laptops; thus, it is compatible with the Microsoft Windows™ software platform. Also, the SEG device needs a simple WiFi connection in order to connect the control glove with the SEG and synchronize the acquired data. The PC, along with the installed GUI application, connects with the SEG device via a simple USB type B cable. Once all components have been connected, the GUI application can be executed smoothly.

The GUI application was developed to function in two different modes. The first mode follows a patient-friendly protocol with five distinct steps. The first step is to register the patient’s code in order to follow the international standards for personal data and the number of exercise cycles that the patient or healthcare professional desires (Figure 7a). Then, as is shown in Figure 7b, the main window appears with the motion bars for each finger, the protocol’s steps, the iterations number, and some generic information about the device, such as the battery condition of the control glove, the connected network, and the version of the application. The next step is the capture of the desired finger motion. Hence, the control glove is activated, and when the motion is completed, the glove transmits the motion data to the SEG device, and the positions are shown on the GUI’s motion bar with white dots. Afterward, the patient performs the same movement without any assistance, and the achieved finger positions are represented on the screen with blue dots. Then, the motion is repeated with the assistance of the SEG in order to achieve the desired positions and activate the muscle memory. This procedure is repeated for the times that have been set at the start of the session. Finally, when the treatment has been completed, the goals and the motion data coupled with the best performance of the patient, in terms of the minimum distance from the goals, are stored on cloud storage with the patient’s code in order to be examined by a healthcare professional and assesses the patient’s progress. On the other hand, the second mode allows the user to operate the SEG manually, setting the desired goals for the motion without the need for the control glove. In this scenario, the graphical environment is the same as the one presented in Figure 7; however, the motion goals are selected with the assistance of a computer mouse. This mode is used for testing purposes and specific movements.

3. Results and Discussion

3.1. Final Design of the SEG Device

In this sub-section, the final design of the overall SEG device is presented, coupled with its components and main technical characteristics. In detail, as is depicted in Figure 8a, the SEG device consists of the main control device and the SEG (1), the control glove (2), the control button (3), the power supply unit (PSU) of the device (4), the USB cable (5), and the air hose (6). Furthermore, Figure 8b shows an illustrative flowchart of the SEG device function. In addition,

Table 1. Technical characteristics of the developed SEG device.

Description	Technical Characteristics
Power Supply	110–220 V (50–60 Hz)
Connectivity	WiFi, USB type B
Air Supply	300–1000 kPa
Operating Air Pressure	200–250 kPa
Circuit Voltage	3.3 V DC
Control Glove Power Supply	5 V DC
Control Glove Battery	Li-Po 3.7 V—550 mAh
Operating Actuators	5

lists the main technical characteristics of the device. It is worth noting that the SEG, along with the main control device, weighs 2.5 kg and the control glove has a mass of 0.1 kg leading to an overall weight of 2.6 kg for the whole SEG device.

3.2. Experimental Procedure and Validation

After the development and assembly of the SEG device, the next step is a basic series of tests in order to verify the functionality and durability of the produced product. The first test concerns the ability of the SEG to produce a sufficient palmar grasping force to accomplish the desired task. According to existing literature [6,35], the ADLs require a palmar grasping force between 10 N to 15 N deriving from the weight of objects of daily living (ODLs), which reach up to 1.5 kg. Therefore, the developed SEG was designed to generate a distal tip force of 7.3 N and 14.5 N at each finger for the operating air pressures of 200 kPa and 250 kPa, respectively. Thus, for a median coefficient of static friction of OBLs at 0.26 [35], which is a conservative estimation, the aforementioned distal tip forces are transformed into palmar grasping forces of almost 1 kg and 1.9 kg for operating air pressures of 200 kPa and 250 KPa, correspondingly. The durability of the SEG device was measured in terms of consecutive full flexion motion cycles (i.e., from 0° to 180°). The layout of the experiment is depicted in the left image of Figure 6c. The frequency of the angular flexion motion was 30 cycles per minute which is a relatively high speed for APLs. Hence, with this configuration, the SEG showed a remarkable performance, as it had undergone more than 40,000 flexion cycles before the failure of one of the actuators due to the fatigue of the silicone rubber compound. It is worth mentioning that in the case of an actuator’s failure, the actuator could be easily replaced. Furthermore, the motion’s speed of the SEG is constant during the device operation and could be changed by the output pressure of the flow regulator. In the context of this study, the nominal motion speed has a value of 90°/s for an output pressure of 250 kPa.

Having proven the functionality and durability of the SEG device, the next process was the proper validation of the device in order to become a commercial product and be ready for clinical trials. Therefore, the designed SEG device underwent a series of tests in order to prove its compliance with health, safety, and environmental protection requirements deriving from the European Union (EU) Directives and to obtain the CE marking. In detail, the tests examined the device for electromagnetic compatibility (EMC test), electrical safety (low voltage directive—LVD test), Restriction of Hazardous Substances (RoHS evaluation), etc. All these tests were performed by an independent contractor certificated by the EU, namely the company LABOR S.A. (European Notified Body 2537). The presented configuration of the developed SEG device successfully passed all the required tests and received the CE mark. Hence, the SEG device is now a commercial product that complies with the EU Directives. However, to expand the application of the device in the health sector, it was essential to evaluate its performance through clinical trials. Towards this direction, a protocol has been developed to compare the SEG device with the standard of care; the protocol was evaluated by the Greek National Organization for Medicines and the Greek National Ethics Committee and received the green light to proceed to clinical trials classifying the device as Category IIa (moderate danger) due to the active motion of the SEG and the existence of electric power in the device.

In order to obtain additional feedback on the usability and efficiency of the developed device (prior to the completion of the official clinical trials), a preliminary evaluation study took place with 30 participants. More specifically, in this study, we used 25 healthy participants aged over 25 years (14 male and 11 female) as a control group, and five patients with limited functionality of their hand recruited following the inclusion criteria described in

Table 2. Profile of the participants for the conducted clinical trials.

Description	Patient Characteristics
Age	>25
Gender	Male and female
Disabled limb	Left or right hand
Rehabilitation stage	Partial hypoxia
Disability medical condition	Monoplegia and hemiplegia
Pathological condition	Cerebrovascular accident
Level of disability	1–3 (Ashworth scale)

. These 30 participants underwent the same experimental procedure and, at the end, answered a questionnaire of 14 questions with a 1–strongly disagree to 5–strongly agree Likert scale in order to document the opinions of the control group and the patients, concerning their user experience of the utilization of the SEG. In addition, five of the healthy people from the control group involved in our experiment were professional experts from the rehabilitation domain who responded, apart from the first questionnaire, to an additional one (Likert scale-based) consisting of five dedicated questions more relevant to the usability, the efficiency, and the applicability of the developed device, as described in Table 2.

The questions of the first questionnaire evaluated the ease to use and the ease of carrying the device as well as the user-friendly graphical environment, ergonomics, functionality, and safety. The questions of the second questionnaire addressed technical issues and appealed to the operators of the device. These questionnaires were anonymously handed out and answered by all the participants. Regarding the experimental procedure, participants experienced a short session of 10 min with the SEG device exercising movements based on the Action Research Arm Test (ARAT) protocol [36] before answering the questionnaires. In detail, four different exercises were performed, namely finger exercise and holding of objects of daily living (cup, pipe, and box), as illustrated in Figure 9. It is worth mentioning that all the examined movements were captured with the control glove and conducted from the SEG with robotic assistance enabled. After the use of the SEG device, the participants answered the questionnaires.

Table 3. The questions of the first questionnaire and the distribution of answers per question.

1st Questionnaire (Control Group and Patients)		Likert Scale (1–Strongly Disagree to 5–Strongly Agree)
		Control Group					Patients
		1	2	3	4	5	1	2	3	4	5
Q101	The use of the device is simple	-	-	4	13	8	-	1	-	2	2
Q102	The use of the device is comfortable	-	-	2	16	7	-	-	1	2	2
Q103	It was easy to learn how to operate the device	-	-	4	15	3	-	-	1	3	1
Q104	I am familiar with the rehabilitation technologies	2	4	4	12	3	-	-	2	2	1
Q105	The device can be integrated into the rehabilitation schedule	-	-	-	13	12	-	-	1	3	1
Q106	The device’s user interface is user-friendly	-	2	4	10	9	-	1	1	3	-
Q107	The device can offer appropriate movement capability	-	2	3	13	7	-	-	2	2	1
Q108	I feel safe while using the device	-	-	-	6	19	-	-	1	1	3
Q109	I feel that the SEG produces sufficient force	-	-	4	5	16	-	-	-	4	1
Q110	The device is lightweight	-	-	-	4	21	-	-	-	4	1
Q111	The visualized data help and guide the procedure	5	5	7	8	-	1	1	1	2	-
Q112	The presentation of data on the GUI is clear	-	4	13	4	4	-	2	1	2	-
Q113	The device has all the expected functionalities	-	4	4	15	2	-	-	2	2	1
Q114	The use of the device is easily learned	-	3	4	15	3	-	-	1	3	1

and

Table 4. The questions of the second questionnaire and the distribution of answers per question.

2nd Questionnaire (Healthcare Professionals)		Likert Scale (1—Strongly Disagree to 5—Strongly Agree)
Q201	The user can have full control of the device’s operation	-	-	1	-	4
Q202	The motion data from the session are clearly visualized	-	-	-	1	4
Q203	The desired data are easy to export in editable format	-	-	2	2	1
Q204	The device offers a reliable library of each patient’s session data	-	-	-	4	1
Q205	The device can contribute to the rehabilitation process	-	-	-	2	3

list the participants’ answers coupled with the questions of the two questionnaires, respectively.

According to the answers to the first questionnaires, the majority of all participants, i.e., 60% (agree or strongly agree), were familiar with rehabilitation technologies, and 100% of the control group and 80% of the patients found easy the integration of the SEG device into the rehabilitation schedule. Furthermore, the participants supported the simplicity of the device in high percentages of 84% for the control group and 80% for the patients, approving the overall design and configuration. In addition, all participants noticed that the SEG device was lightweight and, according to 92% of the control group and 80% of the patients, the device was comfortable. Regarding the mechanics of the SEG device, 80% of the control group and 60% of the patients stated that the device offers appropriate movement capabilities with sufficient force (84% of the control group and 100% of the patients), showing the functionality and efficiency of the pneumatic motion system. Also, only four of 30 participants disagreed with the phrase that the device has all the expected functionalities, highlighting the high functionality of the SEG. One other aspect of the SEG device is its ease to use. Thus, a series of questions were included in the first questionnaire. The first of these questions concerns the easy-to-learn aspect of how to operate the device (Q1₀₃), with 72% of the control group and 80% of the patients agreeing on the device’s ease of learning. It is worth mentioning that a similar question to Q1₀₃ was added, question Q1₁₄, with identical answers operating as a control question and showing the reliability of the answers. The next question of this type focused on the user-friendly aspect of the device, with only three of 30 participants (two from the control group and one from the patient group) declaring the device as non-user-friendly. Moreover, intermixed answers were obtained for the question about the clear presentation of the data on the GUI, with almost 47% of all participants maintaining a neutral opinion about it. The only negative feedback was received for the statement that the visualized data helped and guided the procedure, with only 32% of the control group and 40% of the patients agreeing with the statement. Finally, the vast majority of the participants (almost 97%) felt safe while using the device.

The second questionnaire was answered by five of the healthy participants who were professional experts from the rehabilitation domain and was dedicated to the evaluation of the device from a more technical aspect. For the majority of the questions, the participants gave positive feedback. More specifically, four of five participants stated that the user had full control of the device during the procedure. Furthermore, all participants agreed that during the session, the motion data were clearly visualized and could be easily used for the development of a reliable library for each patient’s data session. In addition, all the professional experts admitted that the device could contribute to the rehabilitation process. Only about the statement regarding the ease of exporting data in editable format 40% of them did not agree.

In general, the results of the questionnaires showed that the majority of participants were satisfied with the usage of the device, as it is simple and easy to learn how it operates, and highlighted the high functionality and reliability of the SEG device. Moreover, the software platform of the device was evaluated by the users as user-friendly and easy to operate. Nevertheless, the questionnaires also showed the weak aspects of the overall product that could be improved in future versions, such as the visualization of the data, especially for non-professional users.

4. Conclusions

In the current study, the design and development of a fully functional soft exoskeleton glove (SEG), along with its software and peripheral hardware, were analytically presented. More specifically, the kinematic mechanism and the bio-inspired design of the employed actuators are described. A novel method of applying a more active mirror therapy was proposed with the introduction of a control glove that dictates the desired motion for the SEG in order to produce a more natural and realistic movement. The hardware configuration of the control glove and the electro-pneumatic control system was analyzed in-depth. Moreover, the basic development strategy for the control software and interaction between the control glove and SEG is presented by utilizing PID controllers driven by the angular motion data from the bend sensors. A user-friendly graphical user interface was developed with proper documentation for patient performance and with respect to the personal data international legislation. The developed SEG device revealed significant performance in terms of durability, with 40,000 motion cycles per actuator and a sufficient palmar grasping force at 19.5 N at maximum operation air pressure. The designed SEG device successfully passed the CE marking tests and has an ongoing certification for the Quality Management Systems standardization for medical devices. Finally, the clinical protocol for the clinical trials was accepted by the corresponding authorities, and the implementation of the protocol was initiated. Additional testing by healthy volunteers (control group) with knowledge and/or experience in rehabilitation and patients was conducted with the use of two questionnaires. The answers of the participants verified the ease to use, ease to carry, user-friendly aspects, the functionality of the SEG device, safety, etc., and highlighted the potential improvement of the SEG’s design to improve the operability of the device.

5. Patent

The research of the current manuscript led to the patent US11141341B2: “System and method for stroke rehabilitation using position feedback based exoskeleton control introduction.” https://patents.google.com/patent/US11141341B2/en (accessed on 26 December 2022).