A comparison of filtered models for dynamic identification of robots

This paper presents a comparison of three models which can be used to identify the minimum dynamic parameters of robots: a dynamic model which depends on the joint acceleration and needs an explicit derivation of the velocity, called the explicit dynamic model; a dynamic model which avoids using explicit acceleration but needs the derivation of a function of the velocity, called the implicit dynamic model; and the energy model, which doesn't need neither acceleration nor implicit velocity derivation. A power model, which is the differential expression of the energy model, is introduced to enlighten the comparison between dynamic and energy models and to improve the filtering of the energy model. Theoretical analysis is carried out from a filtering point of view and clearly shows the differences between the 3 identification models. These results are checked from comparing simulated and experimental identifications of the dynamic parameters of a planar SCARA prototype robot.