2. University of Coimbra
Safe collaborative robot manipulators
2 Main research interest
Collaborative Robotics
Mohammad Safeea
3. University of Coimbra Mohammad Safeea
Safe collaborative robot manipulators3
Safe collaborative robots, what does it mean?
4. University of Coimbra Mohammad Safeea
Safe collaborative robot manipulators4
1. Collaborative robot: The robot shall collaborate with the
human, work with proximity from the human, interact
physically with the humans.
2. Safe robot: it means that robot shall be safe to the human
during the human-robot interaction.
5. University of Coimbra Mohammad Safeea
Safe collaborative robot manipulators5
Main research topics:
Collision avoidance
Precise hand-guiding
Collaboration
Safety
6. University of Coimbra Mohammad Safeea
Safe collaborative robot manipulators6
Collision avoidance
7. University of Coimbra Mohammad Safeea
Robots working in dangerous
environments
Human Robot Collaboration Robots and coworkers are working together
Robots operating
in remote locations.
Three examples
Morphin map
implemented on
Curiosity mars rover.
Collision avoidance in Robotics
Importance of collision avoidance
7
1
23
8. University of Coimbra Mohammad Safeea
Collision avoidance for manipulators
Scientific terms
8
1. Give the ability to perceive
human co-worker, integrating
sensors into a suit that the
human will wear it.
2. Make industrial robots on the
work floor smarter and safer to
humans, the idea is to endow the
robot with advanced motor-skills,
giving them the ability to perform
natural reflexes, with smooth
motions.
In scientific terms,
collision avoidance:
9. University of Coimbra
Attraction and repulsion forces
Attraction forces Repulsion forces
9 Main concept of control strategy
Attraction/repulsion forces
Mohammad Safeea
10. University of Coimbra
• For calculating the repulsion forces, minimum distance shall be calculated.
• For fast minimum distance calculation the robot and the coworker are
represented by capsules.
Collision avoidance for collaborative robotics
Coworker and robot representation
10
Mohammad Safeea
11. University of Coimbra
Sensing of coworker’s configuration and position
IMUs distribution on the body of the coworker
11
The human will be represented
by five capsules, five IMUS will
be used, and one laser scanner
will be used.
Mohammad Safeea
12. University of Coimbra
Sensing of coworker’s configuration and position
Sensors to be utilized
12
The human will be represented
by capsules, the Position and
Configuration of the human will
be captured by sensors.
- The configuration of the
human body will be captured
by IMUs
- The position of the human
body will be captured by SIK
laser sensors, and UWB indoor
positioning sensors.
Mohammad Safeea
13. University of Coimbra
Embedding the collision avoidance in a state machine
Prove of concept, robotic cell simulation
We are implementing the collision avoidance algorithm in a state machine.
13
Mohammad Safeea
17. University of Coimbra
Collision avoidance, case study
Automobile industry
17
Internal side of a car door.
Slots in the internal side of a car door.
Parts to be inserted in the slots.
Case study, how could we use the collision avoidance on
the factory floor.
Mohammad Safeea
18. University of Coimbra
Collision avoidance for collaborative robotics
Case study
18
Case study, how could we use the collision avoidance on
the factory floor.
Assembly order.
Mohammad Safeea
19. University of Coimbra
Collision avoidance for collaborative robotics
Case study
19
The Robotic cell utilized.
Mohammad Safeea
21. University of Coimbra
Some thing to keep in mind
• In the previous tests, the robot is not
preprogrammed to do the reflexes.
• Smart algorithm is used to generate motion
on line.
• As such several obstacle (coworkers) can be
near the robot.
• They can approach the robot from different
direction, and different velocities, the robot
shall avoid collision with them smoothly.
21
Mohammad Safeea
22. University of Coimbra
HRC- Human robot collaboration
Precise hand guiding functionality, for Human Robot Collaboration HRC
22
Precise hand-guiding
Mohammad Safeea
23. University of Coimbra
HRC- Human robot collaboration
Precise hand guiding functionality, for Human Robot Collaboration HRC
23
Why hand guiding:
Confusing??????
Smart-pad is usually used
Keep track of the base frame
orientation
Mohammad Safeea
24. University of Coimbra
HRC- Human robot collaboration
Precise hand guiding functionality, for Human Robot Collaboration HRC
24
The off the shelf KUKA hand guiding
functionality is hard to utilize, and it operates
in joint space, we propose precise hand
guiding controller at end effector level, the
controller is inspired by the properties of
passive mechanical systems. Subjected to
coulomb and viscous friction.
Mohammad Safeea
25. University of Coimbra
25 HRC- Human robot collaboration
Precise hand guiding functionality, for Human Robot Collaboration HRC
Three main motion groups +1 are
defined
Three main motion groups
Mohammad Safeea
26. University of Coimbra
Precise hand guiding, motion groups
26 HRC- Human robot collaboration
Video demonstrator
Mohammad Safeea
27. University of Coimbra
+1 for redundant manipulators,
motion of redundant axes
27 HRC- Human robot collaboration
Video demonstrator
Mohammad Safeea
28. University of Coimbra
Precession of Cartesian linear motion.
28 HRC- Human robot collaboration
Precise hand guiding functionality
Testing
A force is applied in the y
direction, the robot moves
accordingly in the Y direction,
While the coordinates in the X,
Z directions are almost fixed.
From collected data it is found
that the precession in the X
direction is 0.091 mm,
while the average deviation in
the Z direction is 0.092 mm.
Mohammad Safeea
30. University of Coimbra
Comparison between KUKA
off the shelf hand-guiding
and the precise hand
guiding, experimental setup.
30 HRC- Human robot collaboration
Precise hand guiding functionality
Testing
Mohammad Safeea
31. University of Coimbra
Comparison between KUKA off the shelf hand-guiding and
precise hand guiding, Orientation:
31 HRC- Human robot collaboration
Precise hand guiding functionality
Testing
KUKA off the shelf hand-guiding Precise hand-guiding
Mohammad Safeea
The theme, title, of my research is: safe collaborative robot manipulators.
It underlines two sub topics, collaboration and safty.
Industrial robotic manipulators are innately dangerous, while their links posses lots of inertia, and they move with high velocities, any collision between the coworker and the unconscious manipulator could be fatal to the human.
On the issue of safety, I will speak about the collision avoidance.
On the issue of collaboration I will speak about the precision hand guiding.
The code and methods developed are being integrated into the KUKA Sunrise Toolbox, you can download it from my profile at github.
Industrial robotic manipulators are innately dangerous, while their links posses lots of inertia, and they move with high velocities, any collision between the coworker and the unconscious manipulator could be fatal to the human.
Three examples about the importance of collision avoidance algorithm
Control concept, the robot is moving in field of forces, attraction forces that attract the robot toward the goal, repulsion forces that repel the robot away from obstacles.
For calculating the repulsion forces, we need to calculate the minimum distance between robot and obstacle. For fast calculating the minimum distance between the robot and obstacles, they are represented by capsules.
Now we are representing the coworker with five capsules, four for the arms, and one along the torso.
To capture the position and configuration of the coworker, a sick laser sensor, imus and magnetic sensors are being utilized.
Then the collision avoidance controller is embedded into a state machine for performing complicated tasks.