Development of learning algorithm of passive joint for 3R under-actuated robot / Mohd Amiruddin Fikri Yaakob
Position angle analysis by learning algorithm on robotics is extremely important and is widely used as a tool for predictive maintenance to detect faults and mechanical problems. However, in this project position angle analysis was limited to determine position angle for passive joint. Two different...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/15675/1/15675.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Position angle analysis by learning algorithm on robotics is extremely important and is widely used as a tool for predictive maintenance to detect faults and mechanical problems. However, in this project position angle analysis was limited to determine position angle for passive joint. Two different techniques were tested using three rotations (3R) under-actuated robot manipulator. The approach embedded Artificial Neural Network (ANN) algorithm and SIMULINK block diagram. Experiments were conducted to predict an algorithm on position angle measurement either SIMULINK block diagram or program code method applied to three joints; Active 1, Active 2 and Passive respectively. MATLAB software was utilized for data acquisition and analysis for the passive position of 3R under-actuated robot manipulator. The experiment test-rig used in this study was a platform with two (2) DC motor for active (Active 1 and Active 2) joints and rotary digital encoder for acquiring real time output of position angle. Joints of Active 1 and Active 2 were controlled by DC motor and the reference angles were between 0 degree to 45 degree with 5 degree intervals.Overall position angle for passive of both techniques were evaluated and compared.Based on those methods, observations of the correlation between INPUT-OUTPUT relationships have shown positive achievement for positioning of the passive joint of 3R under-actuated robot manipulator. As a conclusion, the results of the experiment on both of the methods have potentially shown relation to the prediction capacity of the algorithm for the 3R under-actuated robot. As a result, ANN of experiment was in acceptable in terms of positioning accuracy and prediction of passive joint. |
|---|