Model-based control design of the compliant robot arm “AirArm”

This paper presents modeling and nonlinear controller design of a 4-axis manipulator arm driven by pneumatic muscle actuators. This arm, named AirArm, was developed within a research project of Festo´s Bionic Learning Network. Since each joint is actuated by two pneumatic muscles, not only the joint position but also its mechanical compliance can be controlled directly. Experimental results show a good performance and dynamics of the 4-axis manipulator arm using the presented nonlinear control concept, which can be separated into a local controller on the joint level and a global MIMO controller on the system level.