Adaptive Integral Inverse Kinematics Control for Lightweight Compliant Manipulators

In this letter, an adaptive to unknown stiffness algorithm for controlling low-cost lightweight compliant manipulators is presented. The proposed strategy is based on the well-known transpose inverse kinematics approach, that has been enhanced with an integral action and an update law for the unknown stiffness of the compliant links, making it valid for soft materials. Moreover, the algorithm is proven to guarantee global task-space regulation of the end effector. This approach has been implemented on a very low-cost robotic manipulator setup (comprised of 4 actuated and 3 flexible links) equipped with a simple Arduino board running at 27 Hz. Notwithstanding, the strategy is capable of achieving a first-order-like response when undisturbed, and recover from overshoots provided by unforeseen impacts, smoothly returning to its nominal behaviour. Moreover, the adaptive capabilities are also used to perform contact tasks, achieving zero steady-state error. The tracking performance and disturbance rejection capabilities are demonstrated with both theoretical and experimental results.