Comparing approaches for actuator redundancy resolution in biarticularly-actuated robot arms

Biarticular actuators - actuators spanning two joints - play a fundamental role in robot arm designs based on the human musculoskeletal actuation structure. Unlike kinematic redundancy, actuator redundancy resulting from biarticular actuation brings advantages such as increased stability, reduced link inertia, and decreased nonlinearity of the end-effector force with respect to the force direction. The way the actuator redundancy is resolved is a fundamental problem, as it strongly characterizes robot arms performance. In this study, the six most significant actuation redundancy resolution approaches in the literature - 1-norm, 2-norm, infinity-norm, phase different control (PDC), nonlinear hase different control (NLPDC), and linear programming (LP) - are analyzed with respect to their design, and experimentally compared with each other using BiWi, a biarticularly actuated and wire-driven robot arm. In addition, an integrated control framework to resolve actuator redundancy maximizing end-effector force and simultaneously minimizing the necessary input torques is proposed.