Identifiability of the Dynamic Parameters of a Class of Parallel Robots in the Presence of Measurement Noise and Modeling Discrepancy

Advanced model-based control schemes and the solution of the direct dynamic problem require accurate knowledge of the dynamic parameters of robotic systems, mainly the inertial properties of the links and the friction parameters at the kinematic joints. A well-known and a very useful tool for their determination is through a dynamic identification process. Normally, in this process, only a subset of the dynamic parameters of a robot, known as base parameters, can be identified. When parameter identification is performed experimentally, not all the aspects of the robot can be modeled in detail. Moreover, measurement variables are affected by noise. These sources of error lead to the fact that not all the base parameters can be properly identified. Therefore, in this paper, the identifiability of the dynamic parameters of a class of parallel robot, in the presence of noise in measurement and discrepancy in modeling, is addressed. The analysis is carried out by means of a simulated robot and over an actual parallel 3-RPS robot.