A Fast New Algorithm for a Robot Neurocontroller Using Inverse QR Decomposition

A new adaptive neural network controller for robots is presented. The controller is based on direct adaptive techniques. Unlike many neural network controllers in the literature, inverse dynamical model evaluation is not required. A numerically robust, computationally efficient processing scheme for neural network weight estimation is described, namely, the inverse QR decomposition (INVQR). The inverse QR decomposition and a weighted recursive least-squares (WRLS) method for neural network weight estimation is derived using Cholesky factorization of the data matrix. The algorithm that performs the efficient INVQR of the underlying space-time data matrix may be implemented in parallel on a triangular array. Furthermore, its systolic architecture is well suited for VLSI implementation. Another important benefit of the INVQR decomposition is that it solves directly for the time-recursive least-squares filter vector, while avoiding the sequential back-substitution step required by the QR decomposition approaches.

[1]  Ali M. S. Zalzala Model-based adaptive neural structures for robotic control , 1996 .

[2]  Alan S. Morris,et al.  Neuro-adaptive control of robotic manipulators , 1993, Robotica.

[3]  C. Cheah,et al.  A model reference learning control scheme for a class of nonlinear systems , 1997 .

[4]  Bin Yang,et al.  Rotation-based RLS algorithms: unified derivations, numerical properties, and parallel implementations , 1992, IEEE Trans. Signal Process..

[5]  A. S. Morris,et al.  Stable and fast neurocontroller for robot arm movement , 1995 .

[6]  Yoshiki Uchikawa,et al.  Trajectory control of robotic manipulators using neural networks , 1991 .

[7]  Luca Maria Gambardella,et al.  On the iterative learning control theory for robotic manipulators , 1988, IEEE J. Robotics Autom..

[8]  Danwei Wang,et al.  Convergence and robustness of a discrete-time learning control scheme for constrained manipulators , 1994, J. Field Robotics.

[9]  S. Shankar Sastry,et al.  Adaptive Control of Mechanical Manipulators , 1987 .

[10]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[11]  Danwei Wang,et al.  Robust motion and force control of constrained manipulators by learning , 1995, Autom..

[12]  Maria D. Miranda,et al.  A hybrid least squares QR-lattice algorithm using a priori errors , 1997, IEEE Trans. Signal Process..

[13]  M. E. Ahmed,et al.  Neural-net-based direct self-tuning control of nonlinear plants , 1997 .

[14]  Filson H. Glanz,et al.  Application of a General Learning Algorithm to the Control of Robotic Manipulators , 1987 .

[15]  Ali M. S. Zalzala,et al.  Neural networks for robotic control : theory and applications , 1996 .

[16]  Maarouf Saad,et al.  Adaptive versus neural adaptive-control : application to robotics , 1994 .

[17]  G. Feng,et al.  Adaptive Tracking Control of Robot Manipulators in Cartesian Space Coordinates , 1991, 1991 American Control Conference.

[18]  Suguru Arimoto,et al.  Learning control for robot tasks under geometric endpoint constraints , 1992, Proceedings 1992 IEEE International Conference on Robotics and Automation.

[19]  Phillip A. Regalia,et al.  Numerical stability properties of a QR-based fast least squares algorithm , 1993, IEEE Trans. Signal Process..

[20]  J. G. McWhirter,et al.  Recursive Least-Squares Minimization Using A Systolic Array , 1983, Optics & Photonics.

[21]  J. G. McWhirter,et al.  Systolic array for recursive least-squares minimisation , 1983 .