방호-임무 수행을 위한 하지 외골격로봇의 자세 안정 제어 운용 프레임워크

초록

Protective-mission tasks involve heavy equipment handling, prolonged static postures, and unexpected disturbances, which lead to biomechanical stress on the lower limbs. While most lower-limb exoskeletons focus on mobility assistance, few address postural stability, and those that do often lack robustness under static or unpredictable conditions. To address this, we propose a variable stiffness virtual spring system-based control framework that enhances postural stability in military-use exoskeletons. The framework estimates the relationship between the CoM and the BOS to monitor balance status, and applies a virtual spring system to the torso when the CoM approaches the BOS boundary. To handle external perturbations, a frequency domain disturbance classification method is employed based on IMU data. A spectral norm derivative of disturbance magnitude, obtained via fast Fourier transform is subsequently used to distinguish involuntary collisions from voluntary user actions. When a collision is detected, the gain of virtual spring system is adjusted to reinforce stabilization. Experiments with an exoskeleton prototype demonstrated improved postural recovery, validating the proposed method’s robustness in mission-oriented task scenarios.

키워드