A New Recursive Instrumental Variables Approach for Robot Identification

Mathieu Brunot (1) , Alexandre Janot (1)

1 DTIS, ONERA, Université de Toulouse [Toulouse]

The work presented in this paper focus on robot identification and presents a method based on the use of instrumental variables (IV). When dealing with en-bloc and offline identification of robots, the instrumental matrix constructed with the inverse dynamic model (IDM) and simulated data obtained from the simulation of the direct dynamic model (DDM). In this paper, a new recursive IV approach relevant for robot identification is presented. The instrumental matrix is constructed with the IDM and the references and their derivatives which are previously filtered by the transfer function of the position closed loop. This new way of building the instrumental matrix avoids the simulation of the DDM and offers some perspectives for online identification and real-time implementation. This recursive IV method termed IDIM-RIV (Inverse Dynamic Identification Model Recursive Instrumental Variables) is experimentally validated on the two degrees-of-freedom SCARA robot. Finally, some hints for real-time implementation are provided.

Mots clés: CLOSED-LOOP IDENTIFICATION; ROBOT IDENTIFICATION; RECURSIVE INSTRUMENTAL VARIABLES

Domaines: Sciences de l'ingénieur [physics]; Physique [physics]; Mathématiques [math]; Informatique [cs]

Mathieu Brunot, Alexandre Janot. A New Recursive Instrumental Variables Approach for Robot Identification. IFAC-PapersOnLine, 2018, 51 (15), pp.132-137. ⟨10.1016/j.ifacol.2018.09.103⟩. ⟨hal-02472764⟩