论文信息 - STOCHASTIC ESTIMATE PCA ALGORITHM FOR PLANAR INDUSTRIAL MANIPULATOR CONTROL

STOCHASTIC ESTIMATE PCA ALGORITHM FOR PLANAR INDUSTRIAL MANIPULATOR CONTROL

In this paper an innovative Stochastic Estimate Algorithm based on Principal Components Analysis (PCA) and on Maximum Likelihood technique, which is applied to estimate nonlinear multi – input – multi - output model (M.I.M.O.) of a manipulator with two link and non-flexible joints, is proposed. As is well known, a problem in mathematical model of a robotic manipulator concerns non-modeled dynamics, for example the static and dynamic friction. A training algorithm is necessary for estimating parameters of a approximate model with linear or nonlinear structure. The proposed algorithm is able to train parameters of a linear digital filter with infinite impulsive answer (I.I.R.). The filter is called "LPC (Last-Principal-Component)", because the solution of the optimization’s problem is the last principal component. Input – Output training set has been opportunely extracted from proportional-derivative (PD) control system of a robotic manipulator which is a centralized control system. Algorithm’s efficiency is validated by digital simulation’s experiments using Matlab 6.5 software. Trajectories for testing PCA algorithm have been implemented in C language using software, hardware and graphic interface of an industrial manipulator.

M. Melluso | T. Raimondi | T. Raimondi | M. Melluso

[1] P. D. Groen. An Introduction to Total Least Squares , 1998, math/9805076.

[2] B. Atal. A model of LPC excitation in terms of eigenvectors of the autocorrelation matrix of the impulse response of the LPC filter , 1989, International Conference on Acoustics, Speech, and Signal Processing,.

[3] S. Van Huffel,et al. A novel neural approach to inverse problems with discontinuities (the GMR neural network) , 2003, Proceedings of the International Joint Conference on Neural Networks, 2003..

[4] I. Jolliffe. Principal Component Analysis , 2002 .

[5] Giansalvo Cirrincione,et al. Direct torque control of induction motors by use of the GMR neural network , 2003, Proceedings of the International Joint Conference on Neural Networks, 2003..

[6] M. G. Jayne,et al. Direct torque control of induction motors , 1995 .

[7] Tommy W. S. Chow,et al. An accelerated recurrent network training algorithm using IIR filter model and recursive least squares method , 1997 .

[8] Michael E. Tipping,et al. Probabilistic Principal Component Analysis , 1999 .

[9] Sanjoy Dasgupta,et al. A Generalization of Principal Components Analysis to the Exponential Family , 2001, NIPS.

[10] F. Girosi,et al. Networks for approximation and learning , 1990, Proc. IEEE.