Figure 5

Overview of the parameter identification procedure under consideration. The figure shows the general approach to parameter identification of industrial manipulators and mechanisms using KPIs. We assume the manipulator has to perform a measurement task on certain TCP positions s requiring a high precision. These positions are considered as KPIs for the parameter identification procedure. The required precision is achieved by combining a (re-)identification of the parameters with the actual work cycle of the manipulator, i.e. whenever the manipulator is idle it can sequentially identify itself. Both the measurement and parameter identification task take place in the same scene. From scene and manipulator model the necessary obstacle descriptions are required to setup the collision avoidance for the OED problem. The OED problem will find an optimal set of configurations q for the optimization problem considering the collision avoidance and the chosen KPI from a given set of initial configurations. Whenever the manipulator performs an identification task, the internal kinematic model is updated with the current optimal parameter values \(\boldsymbol {p}^{\mathrm {true}}\).