Please use this identifier to cite or link to this item:
Full metadata record
|dc.contributor.author||Tiruneh, Habtemariam Alemu||-|
|dc.description.abstract||This dissertation reports results on three different sub-topics under the main theme of GA tuned neural controller for force and position control of robot manipulator: free motion control of robot manipulator with and without artificial neural network compensator (ANN), joint space hybrid force and position control of robot manipulator with neural networks and ANN weight and bias parameters tuning using GA for robot position and force control. Firstly, efficient dynamics models of 6-DOF industrial manipulator are developed using two different approaches: MATLAB/Simulink and MATLAB/SimMechanics. Conventional PID and partitioned PD controllers are designed for free motion control. Neural network auxiliary controller is introduced with PID to counter act uncertainties and disturbances in free motion control. Multilayer feedforward backpropagation artificial neural networks (ANN) with different hidden layer configuration are designed and their performance is tested. The major purpose of this work is to present GA tuned neural controller for joint space hybrid force and position control of robot manipulator. Two separate loops: PID/PPD control law for position control loop and PD control law for force control loop are designed. The method is different from conventional operational-space approach. Genetic algorithm is applied to fined optimal weight and bias values of neural network in off-line. To get better performance, ANN is again further trained in off-line by feedforward backpropagation algorithm using GA tuned weight and bias values as initial parameters. In hybrid force and position control scheme, the complete dynamic model of the manipulator and environment stiffness are considered as approximately known. Through out the work, 6-DOF PUMA 560 robot manipulator is assumed as industrial robot. All the dynamic and kinematics modeling, controller implementation and simulations are described with the dynamic behavior and configuration of PUMA 560.||en_US|
|dc.subject||FORCE AND POSITION CONTROL||en_US|
|dc.title||GA-TUNED NEURO-CONTROLLER FOR FORCE AND POSITION CONTROL OF ROBOT MANIPULATOR||en_US|
|Appears in Collections:||MASTERS' DISSERTATIONS (Electrical Engg)|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.