Bio-Mechatronics Design and Manufacturing of Arm Exoskeleton with Electro-Pneumatic Mechanism for Passive Rehabilitation

Exoskeletons are crucial for providing intensive and consistent rehabilitation over a longer period and may be able to treat the patient without the presence of the therapist compared to manual therapy. This approach allows for frequent treatment reducing several costs. Therefore, this study aimed to examine the passive elbow rehabilitation of lateral epicondylitis patients, usability, and bioinspired design, to develop a mechatronics system with three rehabilitation positions. Regarding the biomechanical fundamentals of the elbow joint and as an engineering sustain, Computer-Aided Design (CAD) was made, consisting of Finite Element Analysis (FEA), anthropometric, ergonomic, and assembly analysis. The results showed that for the three rehabilitation positions, FEA showed von Mises stress less than the elastic modulus by a 10³-factor resulting in no permanent deformation. Position 1, 2, and 3 produced angular displacements of 27°, 16.5° and 31° respectively with a total of 74.5°. An arm exoskeleton for passive rehabilitation of the elbow was developed using a pneumatic cylinder and an AD8832 electromyography (EMG) sensor, capable of detecting the EMG peak point to activate or deactivate the 24 V Arduino relay to flex or extend the elbow based on the positions. A total angular displacement of 74.5° was obtained instead of the simulated version 84.63°, with an error margin of 11.96%. The force during the three rehabilitation positions was 18 N exerted by the air compressor at a 6-bar constant pressure, and due to the use of choke valves.

Arm Exoskeleton; Bio-design; Passive Rehabilitation; Robot-Aided Rehabilitation

A literature review on exoskeletons and robotic rehabilitation showed significant advancements in control and torque estimation. A Model Reference Adaptive Control (MMRAC) has proven effective for controlling soft robotic exoskeletons, enabling both passive and assistive control without the need for additional sensors, and showing robustness against uncertainties (Toro-Ossaba et al., 2024). Additionally, trend analysis showed that all analyzed exoskeletons use flexion/extension movements, with aluminum and 3D printing, as well as PID control algorithms and servomotors (Huamanchahua et al., 2021). The most common applications are assistance and rehabilitation (Cornejo et al., 2021). An adaptive controller based on artificial neural networks demonstrated high accuracy in gesture detection for wheelchair-mounted exoskeletons (Schabron, Desai, and Yihun, 2021). Twisted String Actuators (TSAs) have proven effective in lightweight, customizable assistive devices (Hosseini et al., 2020). Topological optimization of transtibial prosthetic sockets using Finite Element Analysis (FEA) demonstrated improvements in stress performance and weight reduction, although it may also contribute to material fatigue over time (Faadhila, 2022). Compared to manual therapy, exoskeletons have the potential to provide intensive and consistent rehabilitation over a longer period (Lo and Xie, 2012).

Other studies focused on the development and evaluation of innovative medical and rehabilitation devices aimed at enhancing patient outcomes. For instance, a study designed a Transforaminal Lumbar Interbody Fusion (TLIF) spine cage using reverse engineering, followed by simulated compression tests, which demonstrated the capability to withstand the forces typically encountered in spinal fusion surgeries (Norli et al., 2024). Another study addressed the challenge of maintaining motivation in pediatric rehabilitation, particularly for children with Developmental Coordination Disorder (DCD). This study discussed the iterative co-creation of Matti, a pressure-sensitive Tangible User Interface (TUI) for exergaming, which has shown potential as a tool to improve patient engagement and the effectiveness of therapy sessions (Ockerman et al., 2024). These devices can provide autonomous rehabilitation for patients, potentially reducing healthcare costs and increasing treatment frequency. The main focus is on the bio-mechatronics design and manufacturing of an arm exoskeleton with an electro-pneumatic mechanism tailored for passive rehabilitation.

 Passive robotic rehabilitation aims to restore muscle and joint function by assisting the affected limb without requiring active participation from the patient (Juarez et al., 2021). Systems using electromyography (EMG) signals for rehabilitation are typically categorized as active (Tiboni et al., 2018), where movements from a functional limb are mirrored by the affected one (Gull, Bai, and Bak, 2020; Qassim and Hasan, 2020). However, this study focuses on a passive reflex rehabilitation method, where both arms are simultaneously engaged, enhancing recovery through synchronized motion. The classification for the device based on the type of assistance provided is passive reflex (Maciejasz et al., 2014) because it aims to assist the motion of the elbow extremity while mirroring the other arm (Khalid et al., 2023), The proposed exoskeleton assists in elbow movement, a critical function in daily tasks, and is designed to improve rehabilitation outcomes by mirroring the movements of the unaffected limb.

Upper limb exoskeletons are used to rehabilitate patients with various conditions, such as strokes (Cervantes et al., 2024), cerebral palsy (Sandoval et al., 2023), and neuromuscular diseases, helping to regain mobility of the arms, shoulders, hands, and elbows (Lo and Xie, 2012). The elbow is a hinge-type synovial joint located between the humerus with the ulna and radius formed by the humeral trochlea, the spheroidal condyle, the trochlear incision, as well as the head of the radius that allow flexion and extension movements at an angle of up to 170° (Rodriguez et al., 2022). The proximal radioulnar joint allows pronation and supination of the forearm (Molina et al., 2023a). Rotation occurs within a ring formed by the annular ligament and the radial incision of the ulna, and these movements are essential for everyday tasks (Rahlin, 2024).

Mechatronic virtual bio-design and biomechanics are key areas for developing effective arm exoskeletons for passive elbow rehabilitation (Cornejo, Cornejo-Aguilar, and Perales-Villarroel, 2019). Virtual bio-design allows the operation of these devices to be analyzed using computational models, which has led to significant advances in design (Cornejo, Vargas, and Cornejo-Aguilar, 2020). However, challenges such as improving comfort, portability, and efficiency remain. Elbow biomechanics, on the other hand, studies the physical and mechanical principles of joint movements, providing crucial information for the design of exoskeletons that adapt to the anatomy and natural movements of the elbow (Molina et al., 2023b). This study contributes to ongoing efforts by integrating mechatronic virtual bio-design with elbow biomechanics, creating a low-cost, more effective, and personalized exoskeleton for passive elbow rehabilitation. The objective is to enhance the design and performance of exoskeletons through FEA and dynamic simulations, offering a novel approach to rehabilitation technology.
      The manuscript is organized as follows; Section 2 describes materials and methods for the bio-mechatronics design of the arm exoskeleton for passive rehabilitation, including simulation of the control system. Section 3, focuses on the system manufacturing and integration, while Section 4 describes the test and results, focusing on the angular displacement and performance of the system. In Section 5, the manuscript ends with conclusions and further work.

        This started by defining the problem and proposing an innovative solution to the classic rehabilitation methods, with a robot-aided system intended to be accessible and easy to use for patients (Cornejo et al., 2023). To achieve this, the study problem consisted of assessing the passive elbow rehabilitation of lateral epicondylitis patients and clearly understanding the symptoms. As a next phase, the project was subjected to evaluation of the clinical background as well as the usability of bioinspired design, to develop a mechatronics system with the appropriate materials. The specific requirements and constraints were then defined to produce a conceptual design regarding the biomechanical fundamentals of the elbow joint. The material selection consisted of appropriate elements for the exoskeleton and the actuators. As an engineering sustain, Computer-Assisted Design (CAD) was performed, consisting of FEA, anthropometric, and ergonomic, as well as assembly analysis. Consequently, the mechatronics system design, simulation, and manufacturing were prepared. This passive elbow system can be used in rehabilitation and clinical centers, as shown in Figure 1.

2.1. Digital twin
        A digital twin can be described as an integrated multi-physics, multi-scale, probabilistic simulation of a complex product that uses the best available physical models and sensor updates to mirror the life of corresponding twin according to Michael Grieves (Jiang et al., 2021). Although initially focused on industrial applications, digital twins have reached the medical sector, for robot-aided rehabilitation, where the main goal should be relatively simple mechanically. The structure must be easy to put on and training within the limited space, with an additional bio-signals tracking system (Falkowski et al., 2023). Following the digital twin and the design methodology for rehabilitation robots (Martínez and Z.-Avilés, 2020), the mechanical, electronic, control, and programming system for a passive elbow rehabilitation exoskeleton station was developed.

Figure 1 Experiment set-up and apparatus.

2.2.  Free body diagram 
        Initially, the system was encharged to transmit force to replicate the flexion and extension of the elbow, which naturally has a range from 0° to 150° (Martin and Sanchez, 2013). The functional range of movement is from 30° to 130° (Felstead and Ricketts, 2017), hence, a smaller range was considered to not damage the elbow joint, from 30° to 90°. To reach this point, three rehabilitation positions were considered due to the limitation of the pneumatic cylinder at 100 mm axial displacement. Each position produced a 30° angular displacement taking as a reference the position of the joint. As an initial phase, the free body diagram of the system was developed to understand how the system will behave as well as the kinematic and dynamic interactions. Equation (1) describes the geometrical restrictions:

Where  represent the lengths of the link that transmit the force to the revolute joint,  represent the angle formed with a vertical plane as shown in Figure 2. a., by modelling the dynamics of the system using the second law of Newton (Rojas-Moreno, 2001). The linear and rotational motion between the first link and the pneumatic coupling (Figure 2. b.) as well as between the elbow pad and the second link is described in equations (2), (3), and (4) while the variables are presented in Table 1:

Table 1 Description of Variables and Model Parameters

Figure 2 Free body diagram of the system to be developed: a) Geometrical relationships between the links that transmit force, b) Forces interaction on the link and the coupling of the pneumatic cylinder, c) Forces interaction between the link and the revolute joint

2.3.  Mechanical design considerations
        The mechanical design was carried out using ISO 7250-3:2015 (ISO, 2015), which standardizes the dimensions of the elbow to wrist (267 mm), and shoulder to elbow (340 mm). Half of the arm was taken to make the elbow pads considering ergonomics comfort (Schiele and Helm, 2006), with one in the arm of 170 mm, and the other in the forearm of 134 mm. After several simulations, the appropriate link distance was 110 mm to achieve the three rehabilitation positions. For the table support, the proper distance was 50 mm, and when the user feels uncomfortable, regulating the chair height can solve this inconvenience. For the actuator, pneumatic cylinders were considered by offering a high force-to-weight ratio, being able to be stopped at any time without causing injuries, and requiring less maintenance, which is desirable for a potentially mobile rehabilitative exoskeletal system (Karanth P, and Desai, 2022; Burns et al., 2020; Carvalho, Gopura et al., 2016; Zhang et al., 2008). As shown in Figure 3, the hole system can be assembled by the patient putting the table support stable and then connecting the elbow pad, the link mechanism to the pad, and pneumatic cylinder. The parts come with a tolerance of 0.15 mm, allowing a correct coupling, which needs only two ?” screws or the link mechanism. At the same time, the pneumatic cylinder can be introduced by pressure. To make the system accessible, polylactic acid (PLA) was considered, as 3D printing technology can be used to sustain the patient needs in short periods (Demeco et al., 2023), by customizing the designs and maximizing the times of fabrication.

Figure 3 System developed for passive elbow rehabilitation: a) Isometric view, b) Exploded view

2.4.  Finite element analysis
        In the analysis, two groups were considered, one for the elbow pads, and the other for the rehabilitation station. A previous study developed human arm parameters (Speich, Shao, and Goldfarb, 2005), to simulate the weight distribution of the forearm and arm using FEA. As a reference point for the studies, the base of each elbow pad was considered a fixed geometry. A normal force of 150 N was applied to the face in contact with the arm in the elbow pads, by taking a median of 10 kg for the human arm mass, and a scale factor of 1.5 was used for the analysis. Furthermore, the material selected was ABS from the SolidWorks standard library, with an elastic module of 2 GPa. The maximum von Mises stress for elbow pad 1 was 56.234 N/m², and for elbow pad 2, it was 331.985 N/m². Under the material elastic limit, there is no permanent deformation.

         The second group of FEM studies was made to the table support and the linkages. Figure 4a shows that l₁ receives a von Mises stress of 135 MPa in the bolted joint, hence, a ?” screw was considered as the maximum shear stress of 145 MPa according to the ‘Structural Design of Stainless Steel’ (BSSA, 2001). For the table support, Figure 4b shows a maximum von Mises stress of 8.567 MPa for the elbow joint, while Figure 4c shows a maximum of 6.287 MPa for the bolted joint in the pneumatic cylinder coupling. Figure 4d shows a maximum von Miss stress of 20 MPa in the joint to the elbow pad. However, these values are less than the elastic module by a 10³-factor.