Summary
This proposal addresses robotic manipulation planning for human-robot collaboration during manufacturing. My objective is to develop a planning framework which will enable a team of robots to grasp, move and position manufacturing parts (e.g. planks of wood) such that a human can execute sequential forceful manufacturing operations (e.g. drilling, cutting) to build a product (e.g. a wooden table).
The overall objective is divided into three components: First, I will develop a planning algorithm which, given the description of a manufacturing task, plans the actions of all robots in a human-robot team to perform the task. Second, I will develop probabilistic models of human interaction to be used by the planner. This model will include (i) an action model that assigns probabilities to different manufacturing operations (e.g. drilling a hole vs. cutting a piece off) as the next actions the human intends to do; (ii) a geometric model that assigns probabilities to human body postures; and (iii) a force model that assigns probabilities to force vectors as the predicted operational forces. Third, I will build a real robotic system to perform experiments and test my algorithm's capabilities. This system will consist of at least three robot manipulators.
This fellowship will enable me to add a completely new human dimension to my planning research. I will work with Prof. Tony Cohn (supervisor) who is a world-leading expert in human activity recognition and prediction - a critical skill for the human-robot collaboration problem I intend to solve. From him and his group, I will receive training on tracking/predicting human posture and recognizing/predicting human activities using vision and point-cloud data. I will then integrate these tracking and prediction methods into a robotic planning framework to enable human-robot collaborative operations.
This fellowship will help me to attain a permanent academic position and to become a leading researcher in robotic manipulation.
The overall objective is divided into three components: First, I will develop a planning algorithm which, given the description of a manufacturing task, plans the actions of all robots in a human-robot team to perform the task. Second, I will develop probabilistic models of human interaction to be used by the planner. This model will include (i) an action model that assigns probabilities to different manufacturing operations (e.g. drilling a hole vs. cutting a piece off) as the next actions the human intends to do; (ii) a geometric model that assigns probabilities to human body postures; and (iii) a force model that assigns probabilities to force vectors as the predicted operational forces. Third, I will build a real robotic system to perform experiments and test my algorithm's capabilities. This system will consist of at least three robot manipulators.
This fellowship will enable me to add a completely new human dimension to my planning research. I will work with Prof. Tony Cohn (supervisor) who is a world-leading expert in human activity recognition and prediction - a critical skill for the human-robot collaboration problem I intend to solve. From him and his group, I will receive training on tracking/predicting human posture and recognizing/predicting human activities using vision and point-cloud data. I will then integrate these tracking and prediction methods into a robotic planning framework to enable human-robot collaborative operations.
This fellowship will help me to attain a permanent academic position and to become a leading researcher in robotic manipulation.
Unfold all
/
Fold all
More information & hyperlinks
| Web resources: | https://cordis.europa.eu/project/id/746143 |
| Start date: | 01-05-2017 |
| End date: | 30-04-2019 |
| Total budget - Public funding: | 195 454,80 Euro - 195 454,00 Euro |
Cordis data
Original description
This proposal addresses robotic manipulation planning for human-robot collaboration during manufacturing. My objective is to develop a planning framework which will enable a team of robots to grasp, move and position manufacturing parts (e.g. planks of wood) such that a human can execute sequential forceful manufacturing operations (e.g. drilling, cutting) to build a product (e.g. a wooden table).The overall objective is divided into three components: First, I will develop a planning algorithm which, given the description of a manufacturing task, plans the actions of all robots in a human-robot team to perform the task. Second, I will develop probabilistic models of human interaction to be used by the planner. This model will include (i) an action model that assigns probabilities to different manufacturing operations (e.g. drilling a hole vs. cutting a piece off) as the next actions the human intends to do; (ii) a geometric model that assigns probabilities to human body postures; and (iii) a force model that assigns probabilities to force vectors as the predicted operational forces. Third, I will build a real robotic system to perform experiments and test my algorithm's capabilities. This system will consist of at least three robot manipulators.
This fellowship will enable me to add a completely new human dimension to my planning research. I will work with Prof. Tony Cohn (supervisor) who is a world-leading expert in human activity recognition and prediction - a critical skill for the human-robot collaboration problem I intend to solve. From him and his group, I will receive training on tracking/predicting human posture and recognizing/predicting human activities using vision and point-cloud data. I will then integrate these tracking and prediction methods into a robotic planning framework to enable human-robot collaborative operations.
This fellowship will help me to attain a permanent academic position and to become a leading researcher in robotic manipulation.
Status
CLOSEDCall topic
MSCA-IF-2016Update Date
28-04-2024
Images
No images available.
Geographical location(s)
