Utilisation de polynômes de Tchebychev pour l'identification de modèles à temps continu de robots

In this paper, a presentation of Tchebychev polynomials' properties and a formulation derived using these polynomials are performed. The identification method presented consists in applying a linear transformation to the set of equations of the process. For this, the mathematical operator used is based on a decomposition of the signal in a basis of orthogonal polynomials. We show that this projection acts as a bandpass filter, so that only one operation is required for the data preprocessing before the identification, namely the choice of the Chebyshev polynomials basis size. An experimental closed-loop identification of a 2-axis robot with this method is performed in a final section. The results are compared and analyzed with other techniques used in robotics.