Experiments with deep visual servoing for connector insertion
Deep learning has allowed for significant progress to be made in the world of robotics. In todays world, its applications can be seen in many fields, from medical to hospitality and many others [1]. This paper investigates the concept of deep learning and its application in robotic assembly. Buildin...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/138962 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Deep learning has allowed for significant progress to be made in the world of robotics. In todays world, its applications can be seen in many fields, from medical to hospitality and many others [1]. This paper investigates the concept of deep learning and its application in robotic assembly. Building upon a past final year project, this paper also investigates the different possible combinations of features of deep learning to produce a sub-millimetre, highly accurate neural network for robotic assembly for better generalisation to new connectors and environments through various experiments. Through various rounds of experimentation, this project was able to prove that the novel architecture by the previous final year project is able to generalise well to unseen connectors with little training. Translational errors are kept below 0.5 mm in the x, y and z directions while rotational errors in the roll, pitch and yaw directions are below 0.4◦.
. |
|---|