Parameter Identification of the KUKA LBR iiwa Robot Including Constraints on Physical Feasibility

Abstract The newly released KUKA LBR iiwa 14 R820 robot stands for intelligent industrial work assistant (iiwa) and is, like its predecessor LBR IV, equipped with torque sensors in each joint, and can be controlled through a real-time interface. Although the dynamic model of the robot is not published by the manufacturer, its knowledge is indispensable for simulation and control based on the system model. This paper presents the identification of the minimal set of base parameters, as well as a consistent set of physical parameters for a rigid-link model of the KUKA LBR iiwa 14 R820 robot, including friction. The experiments on the robot are conducted based on optimized excitation trajectories. The physical parameters, which are required for stable dynamic simulations, are identified by solving a nonlinear optimization problem, where constraints are included to ensure physical feasibilty. A validation and cross-validation in simulation and experiments show a very accurate representation of the robot’s dynamics by the resulting models. As a result, both sets of identified parameters are given.