Identification of parameters and control of robotic manipulators using artificial neural networks

Aggarwal, Rajnish (1995) Identification of parameters and control of robotic manipulators using artificial neural networks. Masters thesis, Memorial University of Newfoundland.

[English] PDF (Migrated (PDF/A Conversion) from original format: (application/pdf)) - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (14MB) [English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. (Original Version)

Abstract

In the present work the estimation of dynamic parameters and the trajectory control of a two link planar manipulator is carried out. The derivation of system equations involve the kinematic parameters such as joint positions, velocities and accelerations. For both the dynamic parameter estimation and the trajectory control an Artificial Neural Networks method called, Linear Programming (LP)-Neuro Method, is used. In this algorithm, the weights are obtained by a combination of linear programming having a sparse coefficient matrix and a single variable nonlinear optimization routine. -- The training set values required for parameter estimation are generated by the Iterative Newton-Euler Dynamics Algorithm. The Artificial Neural Network is trained to predict the dynamic parameters. The values of the forces and torques are recomputed based on the estimated dynamic parameters. This method is useful for the on line parameter estimation of the manipulators having odd mass distribution, which is the case in actual practice. -- For the control problem non-linear optimization method is used to evaluate the gain parameters required for the manipulator to follow the desired trajectory. Then the Linear Programming (LP)-Neuro Method is used to obtain the weight matrix which relates the input (joint positions and velocities) and the output (gain parameters). This weight matrix, for each point along the trajectory, can be used on-line to evaluate the gain parameters thereby eliminating the time consuming calculations at each point along the trajectory. Finally, the effect of the variation of the maximum tangential velocity and general control law are studied.

Actions (login required)

View Item