Reinforcement learning-based target grasping for low cost robotic arm
This dissertation focuses on grasping targets by robotic arms based on deep reinforcement learning. Theoretically, it goes into the application of the most advanced algorithms of reinforcement learning, particularly the Soft Actor-Critic algorithm. It will choose Stable Baselines3 for many advant...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis-Master by Coursework |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/180764 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | This dissertation focuses on grasping targets by robotic arms based on deep
reinforcement learning. Theoretically, it goes into the application of the most
advanced algorithms of reinforcement learning, particularly the Soft Actor-Critic
algorithm. It will choose Stable Baselines3 for many advantages: verified set
of algorithms, compatibility with PyTorch, and community support. The study
has also been conducted with detailed analysis in terms of the structure of the
robotic arm through juxtaposition of traditional control methods against reinforcement
learning algorithms to gain a deeper understanding as to the various
ways of controlling a robotic arm. |
|---|