Control Strategy for Shoulder-SideWINDER With Kinematics, Load Estimation, and Friction Compensation: Preliminary Validation

Abstract

In industrial environments, the control of arm assistive exoskeletons presents significant challenges due to complex dynamic movements. We propose a high-level controller for the Shoulder-sideWINDER that assists dynamic arm movement during a variety of tasks, based on mechanical characteristics of a novel joint alignment mechanism. The high-level controller consists of control units related to four main considerations: (i) Shoulder-sideWINDER Force Model, (ii) Friction Compensator, which determines the structurally required torque, (iii) Arm Kinematics Estimator, which estimates the required torque according to arm postures using a simple IMU-based approach, (iv) Load Estimator, which estimates muscle load and the required torque caused by external weights and also determines the force engagement. To validate the controller, a right arm system emulated three different postures. One participant operated the Load Estimator by lifting and lowering a 5 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$kg$</tex-math></inline-formula> weight five times with a 3-second interval for each posture. Although, when reducing the torque, a force transmission delay and a residual force were caused by friction, the result shows that the controller estimated correctly and generated the target required torque with an error of less than 6.6%. Finally, we verified that the proposed control strategy is simple but effective in generating the required torque based on four main considerations, and identified the direction for further improvement of the friction compensator that enhances comfort as well as control performance.