Human-Centered Design of a Passive Lower-Limb Exoskeleton Integrated with Mixed Reality for Assistive Strength Training

Sports-related lower-extremity musculoskeletal injuries often lead to muscle weakness, joint instability, and restricted mobility during recovery. This study presents a human-centered rehabilitation system that integrates a passive lower-limb exoskeleton with a mixed-reality training simulator to support static and low-intensity dynamic strength training. Following a human-centered design process, user needs and rehabilitation requirements concerning comfort, adjustability, safety, and training feedback were translated into system specifications and refined through iterative prototyping. The exoskeleton comprises a flexible body-worn interface, a rigid lower-limb support frame, gear-based hip and knee modules, and adjustable compliant joints. A coaxial torsion-spring mechanism with adjustable preload regulates passive assistance, joint neutral position, and range of motion without external actuation, while nested hip-joint components improve anatomical alignment and accommodate variations in body dimensions and movement patterns. The mixed-reality simulator provides visual and auditory guidance for movement demonstration, rhythm regulation, task progression, and performance feedback, thereby establishing an integrated exoskeleton–mixed-reality training framework. The two preliminary evaluations were conducted. First, an sEMG-based pilot experiment with one healthy adult male participant compared a unilateral lower-limb stretching task with and without exoskeleton assistance. Surface electromyography signals from four lower-limb muscles were quantified using mean absolute value (MAV) and root mean square (RMS), two amplitude-based indicators of muscle activation. Exoskeleton assistance increased MAV by 20.54%–44.24% and RMS by 28.05%–50.43%, indicating enhanced target-muscle activation during stretching. Second, 20 participants evaluated the integrated exoskeleton–mixed-reality system while performing three static training tasks. Mean scores for perceived effectiveness, ease of use, and comfort were 6.28, 5.56, and 5.92 out of 7, respectively, with acceptable questionnaire reliability (Cronbach’s α = 0.793–0.820). These findings provide preliminary objective and user-centered evidence for the feasibility and acceptability of the proposed system for assistive lower-limb strength training.