Modeling, parameter identification and model-based control of a lightweight robotic manipulator

Nowadays many robotic tasks require close and compliant tracking of desired positions, paths, or forces. To achieve these goals, model-based control schemes provide a possible solution, allowing to directly consider the nonlinear dynamics. One of the key challenges, however, is the derivation of suitable models, which allow sufficiently fast evaluation, as well as the parameterization of these models based on measurements. In this work we outline and review a structured approach for model-based controller design for robots. In a first step we derive suitable models for multi-link robots. In a second step we review how such models can be parameterized and how optimal identification experiments can be designed. Based on the model we then derive a simple model based controller to experimentally validate the results considering a lightweight robot. The single steps of the derivation and controller design are supported by a newly developed freely available model toolbox for the considered lightweight robot.