Advanced Teleoperation with Haptic Feedback
(2022)Department of Automatic Control
- Abstract
- As the world moves more and more towards automation of repetitive and dangerous tasks there are still many tasks that primarily involve interaction with an object that are still carried out by humans. This is in many ways due to difficulties in programming the robot to achieve similar performance. The most natural way of transferring these skills to the robot is by demonstration. This can be achieved by the use of haptic feedback to allow the operator to feel the robotic system during the demonstration.
This master thesis seeks to implement and improve master-slave coordination using virtual constraints for a redundant dual-arm haptic interface. This is done by implementing a general-purpose algorithm compatible with different types of 6... (More) - As the world moves more and more towards automation of repetitive and dangerous tasks there are still many tasks that primarily involve interaction with an object that are still carried out by humans. This is in many ways due to difficulties in programming the robot to achieve similar performance. The most natural way of transferring these skills to the robot is by demonstration. This can be achieved by the use of haptic feedback to allow the operator to feel the robotic system during the demonstration.
This master thesis seeks to implement and improve master-slave coordination using virtual constraints for a redundant dual-arm haptic interface. This is done by implementing a general-purpose algorithm compatible with different types of 6 degrees of freedom robots in Python. For this project a UR5e robot was used as well as a second virtual clone of the UR5e robot to serve as master and slave arm respectively.
The virtual second UR5e robot was simulated using both Maplesim and inverse dynamics. The inclusion of force-torque (F/T) sensors was implemented and tested.
Free-space motion where position and orientation offsets are retained as well as collision with a soft surface resulting in a feeling of force upon interaction with a soft object were achieved. The singularity-free operation produced consistent and intuitive results, with the inclusion of a teaching handle to improve adjustments. However, low sampling frequency of the implemented algorithm resulted in delays that negatively impacted the translational and rotational error between the end effectors of these robots, as well as not reaching a realistic and intuitive feeling of force upon interaction with a soft object. These problems were not solved due to lack of time.
The impact of the implemented F/T sensor was investigated, but due to the limited time left was this not enough to reach a definitive conclusion.
The implemented algorithm functions well during slower movements, with room for improvements in the optimization of calculation times and further investigation of the impact of the implemented F/T sensor. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9075135
- author
- Ragnarsson, Eric
- supervisor
- organization
- year
- 2022
- type
- H3 - Professional qualifications (4 Years - )
- subject
- report number
- TFRT-6153
- other publication id
- 0280-5316
- language
- English
- id
- 9075135
- date added to LUP
- 2022-02-10 11:50:06
- date last changed
- 2022-02-10 11:50:06
@misc{9075135,
abstract = {{As the world moves more and more towards automation of repetitive and dangerous tasks there are still many tasks that primarily involve interaction with an object that are still carried out by humans. This is in many ways due to difficulties in programming the robot to achieve similar performance. The most natural way of transferring these skills to the robot is by demonstration. This can be achieved by the use of haptic feedback to allow the operator to feel the robotic system during the demonstration.
This master thesis seeks to implement and improve master-slave coordination using virtual constraints for a redundant dual-arm haptic interface. This is done by implementing a general-purpose algorithm compatible with different types of 6 degrees of freedom robots in Python. For this project a UR5e robot was used as well as a second virtual clone of the UR5e robot to serve as master and slave arm respectively.
The virtual second UR5e robot was simulated using both Maplesim and inverse dynamics. The inclusion of force-torque (F/T) sensors was implemented and tested.
Free-space motion where position and orientation offsets are retained as well as collision with a soft surface resulting in a feeling of force upon interaction with a soft object were achieved. The singularity-free operation produced consistent and intuitive results, with the inclusion of a teaching handle to improve adjustments. However, low sampling frequency of the implemented algorithm resulted in delays that negatively impacted the translational and rotational error between the end effectors of these robots, as well as not reaching a realistic and intuitive feeling of force upon interaction with a soft object. These problems were not solved due to lack of time.
The impact of the implemented F/T sensor was investigated, but due to the limited time left was this not enough to reach a definitive conclusion.
The implemented algorithm functions well during slower movements, with room for improvements in the optimization of calculation times and further investigation of the impact of the implemented F/T sensor.}},
author = {{Ragnarsson, Eric}},
language = {{eng}},
note = {{Student Paper}},
title = {{Advanced Teleoperation with Haptic Feedback}},
year = {{2022}},
}