Design of a Self-Aligning Knee Exoskeleton for Minimizing Parasitic Forces During Assisted Ambulation

One major challenge in the application of exoskeletons is the misalignment between the biological and mechanical joint axes during use, which arises from human–robot kinematic mismatches and variability in wearing conditions. As a result, parasitic forces are generated at the physical human–robot interface and lead to significant relative sliding distance (RSD) between the exoskeleton and the human leg due to constrained fit. To address this issue, we propose a novel kinematically adaptive, self-aligning knee exoskeleton that eliminates sagittal-plane misalignment to reduce parasitic forces and removes the need for prealignment before use. Both platform and human–subject experiments were performed to evaluate its performance. The platform tests demonstrate that the exoskeleton can transmit effective pure assistive torque without introducing additional parasitic forces at the pHRI. Human–subject tests show that the exoskeleton maintains low RSD and parasitic forces across walking speeds up to 1.8 m/s and assistance levels up to 16 N$\cdot$m; although both increased with speed and assistance level, peak differences remained below 6 mm and 8 N, respectively. During stand-to-sit transitions under constant assistive torques (up to 8 Nm) at large knee flexion angles, similarly low RSD and parasitic forces were observed, with torque transmission efficiency exceeding 87%.