Elbow joint rehabilitation presents a formidable challenge, underscored by the joint’s complex biomechanics and high vulnerability to injuries and degenerative conditions. Despite the advancements in rehabilitative technology, current solutions such as rigid exoskeletons often fall short in providing the precision, flexibility, and customization needed for effective treatment. Although
... [Show full abstract] traditional robotic aids, such as rigid exoskeletons, help recover, they lack in providing sufficient flexibility, comfort, and easy customization with no need for complicated calculation and complex design considerations. The introduction of soft pneumatic muscles marks a significant development in the rehabilitation technologies field, offering distinct advantages and unique challenges when compared to conventional rigid systems. These flexible actuators closely mimic the elasticity of biological tissues, improving safety and interaction between humans and machines. Designed for individualized therapy, its versatility allows application in various rehabilitation scenarios, from clinical settings to home settings. The novelty of this approach lies in the development of biomechanically-compliant soft pneumatic muscles optimized for precise rotational control of the elbow joint, coupled with an advanced deep learning-based motion tracking system. This design overcomes limitations in force control, stability, and pressure requirements found in existing pneumatic-based systems, improving the safety and efficacy of elbow rehabilitation. In this study, the design, fabrication and systematic evaluation of a soft pneumatic muscle for elbow rehabilitation are presented. The device is designed to closely simulate the complex biomechanical movements of the elbow, with a primary focus on the rotational motions that are essential for controlling flexion and extension, as well as positioning the wrist during grasping tasks. Through the integration of precise geometric parameters, the actuator is capable of controlled flexion and extension, reflecting the natural kinematics of the elbow. Employing a rigorous methodology, the research integrates finite element analysis with empirical testing to refine the actuator’s performance. Under varying air pressures, the soft muscle demonstrated remarkable deformation along the X-axis (10–150 mm) and the Y-axis, indicative of its symmetrical rotational behavior, while maintaining minimal elongation along the Z-axis (0.003 mm max), and proper lifiting force under a maximum wight of 470 gm. highlighting the stability and targeted response of the device to pneumatic actuation. A specialized experimental apparatus comprising a 3D environment, a pneumatic circuit, a LabVIEW-based control system, and a deep learning algorithm was developed for accurate position estimation. The algorithm achieved a high predictive accuracy of 99.8% in spatial coordination tracking, indicating the precision of the system in monitoring and controlling the actuator’s motion.