A generalized recognition of single ended contact formations for use in automated assemble operations /
Robots are preferred over any other form of automated hics.
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1998.
|
| Subjects: |
| Summary: | Robots are preferred over any other form of automated hics. machines for assembly tasks due to their capability of being programmed to perform a variety of tasks. However in the present day industries, the turnaround time for new designs has dramatically reduced. Therefore the need for robots which can adapt teaching and programming to new situations is strongly felt. This is especially true in tasks such as assembly operations, which involve the robot making frequent contact with its environment. This research addresses the problems that arise due to small changes in the work settings after the system has been programmed or trained. In an industry setting. it is very likely that changes such as orientation and translation of the grasped object with respect to the robot axes can occur due to many unforeseen causes. The research here is focused on generalizing a Hybrid Control System, in which an assembly skill is described as a sequence of qualitative states and the desired transition between the states. In this case, the qualitative state takes the form of a single-ended contact formation, which describes how a grasped object touches its environment. Skill acquisition involves learning the sequence of qualitative states, the transition between those states, and the mapping from the sensor signals to the qualitative states. We discuss impact of changes in the orientation and the position of the grasped object with respect to the robot axes on the recognition of these qualitative states. We also propose a method of decreasing the performance degradation caused by this orientation change in recognition of these qualitative states, by adapting to the new situation with as little retraining as possible. Experimental results are presented which illustrate and validate the approach. |
|---|---|
| Item Description: | Vita. "Major subject: Mechanical Engineering". |
| Physical Description: | x, 53 leaves : illustrations ; 28 cm. |
| Bibliography: | Includes bibliographical references (leaves 50-52). |