interaction control of collaborative robots for adaptation ... · −the impedance/force parameters...
TRANSCRIPT
Interaction Control of Collaborative Robots for Adaptation to Task and Human Factors
Arash Ajoudani
Fondazione Istituto Italiano di Tecnologia [email protected]
HRI2 Research
Assistive Robotics / Power augmentation
Physical Human-Robot & Robot-Robot Interaction/Cooperation
Teleoperation/Teleimpedance Human Motor Control Understanding for
Robotic Manipulation
Fondazione Istituto Italiano di Tecnologia [email protected]
Robots with advanced physical interaction capabilities: Integration to human environments
Laghi et al., IJRR 2018
Peternel et al., RAL 2016
Ajoudani et al., Humanoids 2014
Fondazione Istituto Italiano di Tecnologia [email protected]
Robots, Humans, and the Environment
H
R
E
Fondazione Istituto Italiano di Tecnologia [email protected]
Robots, Humans, and the Environment
H
R
E
Fondazione Istituto Italiano di Tecnologia [email protected]
Robot interaction with the environment
− How to tune robot control parameters (i.e., impedance/force gains)?
− The impedance/force parameters are usually pre-defined by
programmers
− This will limit the interaction capacity of the robots!
Fondazione Istituto Italiano di Tecnologia [email protected]
Approach 1: Replicating human physical interaction behavior
S. Fani et al., IEEE RAM 2018
A. Ajoudani et al., IJRR 2017
x 1.5
Fondazione Istituto Italiano di Tecnologia [email protected]
Fondazione Istituto Italiano di Tecnologia [email protected]
Adaptation to external disturbances
A. Scibilia et al., Humanoids 2018
Fondazione Istituto Italiano di Tecnologia [email protected]
Approach 2:Robots with interaction autonomy
− Self-tuning of the impedance gains/interaction force based on the
robot sensory feedback
− How to distinguish good interaction from external disturbances?
− What sensory systems are required to achieve robot interaction
autonomy?
See Pietro’s presentation on Thursday at 9:00 in Room 2.L2 (Control I)
Fondazione Istituto Italiano di Tecnologia [email protected]
Towards robot interaction autonomy:Explore, Identify, and Interact
Kawato, Mitsuo. "Internal models for motor control and trajectory planning." Current opinion in neurobiology 9.6 (1999): 718-727.
Fondazione Istituto Italiano di Tecnologia [email protected]
Towards robot interaction autonomy:Explore, Identify, and Interact
P. Balatti et al., RAL’19, under review
Fondazione Istituto Italiano di Tecnologia [email protected]
Robots, Humans, and the Environment
H
R
E
Fondazione Istituto Italiano di Tecnologia [email protected]
Human Robot Collaboration: SOA
F. Flacco et al., ICRA 2012
Fondazione Istituto Italiano di Tecnologia [email protected]
Can CoBots physically assist humans instead of avoiding them?
Fondazione Istituto Italiano di Tecnologia [email protected]
Physical Overloading Reduction
Fondazione Istituto Italiano di Tecnologia [email protected]
Physical Overloading Reduction
Fondazione Istituto Italiano di Tecnologia [email protected]
Dynamics of interaction
𝑧𝑥
𝑪𝑃𝚺𝑾
𝚺𝟎
𝑪𝑃𝒇𝑤𝑜
𝒇𝑤𝑡
𝒇ℎ∆𝜏
Without interaction forcesWith interaction forces
∆𝜏
∆𝜏
W. Kim et al., RAL 2017
Fondazione Istituto Italiano di Tecnologia [email protected]
Ergonomic Human Robot Collaboration
𝑧𝑥
𝑪𝑃𝚺𝑾
𝚺𝟎
Offline OnlineOffline
WarningShoulder
Hip
Knee∆𝜏 ↓
Moving to reduce
the overloading
joint torque
𝜹𝒑
𝑪𝑃 𝑪𝑃𝒇𝑤𝑜 𝒇𝑤𝑡
𝒇ℎ∆𝜏
Without
interaction
forces
With
interaction
forces
∆𝜏
∆𝜏
Fondazione Istituto Italiano di Tecnologia [email protected]
SESC: Human CoM Model
𝑧𝑥
𝑪𝑃𝚺𝑾
𝚺𝟎
OfflineOffline
Human CoP model using statically equivalent serial chain (SESC)
S. Cotton, et al., ASME (2009).
Identification the body parameters
p : number of posture
See Wansoo’s presentation on Thursday at 18:00 in Room 2. R1 (Haptics)
Fondazione Istituto Italiano di Tecnologia [email protected]
Overloading Joint Torque
CO
Px
RAL 2017, ICORR 2017
Fondazione Istituto Italiano di Tecnologia [email protected]
Human-Robot Load-Sharing
W. Kim, J. Lee, L. Peternel, N. Tsagarakis and A. Ajoudani, "Anticipatory Robot Assistance for the Prevention of Human Static Joint Overloading in Human-RobotCollaboration", IEEE Robotics and Automation Letters (RA-L), 2017.
Fondazione Istituto Italiano di Tecnologia [email protected]
Ergonomic Human-Robot Co-Manipulation
Modified SESC Model
Dynamic CoP
Fondazione Istituto Italiano di Tecnologia [email protected]
Ergonomic Human-Robot Co-Manipulation
L. Peternel et al., Humanoids 2017
W. Kim et al., IEEE Transactions on man machine systems , under review
Fondazione Istituto Italiano di Tecnologia [email protected]
Ergonomic Handover
Fondazione Istituto Italiano di Tecnologia [email protected]
Human-Robot Handover
L. Peternel et al., Humanoids 2017 W. Kim et al., IEEE Transactions on man machine systems , under review
Fondazione Istituto Italiano di Tecnologia [email protected]
Replacing wearable suit with vision
Detection of human in 2D image and then reconstructing 3D information using disparity map by applying with some biophysical constraints (OpenPose model)
Fondazione Istituto Italiano di Tecnologia [email protected]
First Ergonomic and reconfigurable HRC Framework KUKA Innovation Challange 2018
Wansoo Kim
Marta Lorenzini
Pietro Balatti
Phuong Nguyen
x 1.5
Fondazione Istituto Italiano di Tecnologia [email protected]
KUKA Innovation Winner
Prize: 20K Euros
Fondazione Istituto Italiano di Tecnologia [email protected]
Physical Fatigue Management
IEEE Transactions on Neural Systems & Rehabilitation, 2017 / RAL 2017 / IROS’16
Fondazione Istituto Italiano di Tecnologia [email protected]
Physical Fatigue Management
Humanoids 2016 (best paper award finalist) / Autonomous Robots, 2017
See Luka’s presentation on Tuesday at 14:30 in Room 1.L3 (HRI II)
Fondazione Istituto Italiano di Tecnologia [email protected]
Conclusion
−The principles behind human physical interaction can help us achieve similar performances in robots.
−Use of sensory systems to analyse human factors during interaction.
−Adaptation to humans does not necessarily mean avoiding them (coexistence vs collaboration)
−Collaborative robots can do more than repetitive tasks!
Fondazione Istituto Italiano di Tecnologia [email protected]
HRI2 paper presentations at IROS
Control of grasp stiffness through CMS and CDS control- See Virginia ‘s presentation on Wednesday at 12:30 in Room 2.R1 (Grasping III)
Ergonomic Feedback Interfaces for Human-Robot Collaboration- See Wansoo’s presentation on Thursday at 9:00 in Room 2.L2 (Control I)
Human fatigue modeling in Human-Robot Collaboration- See Luka’s presentation Tuesday at 14:30 in Room 1.L3 (HRI II)
Auto-tuning Impedance control- See Pietro’s presentation on Thursday at 9:00 in Room 2.L2 (Control I)
Robot Fatigue modeling and management- See Luka’s presentation on Wednesday at 9:0 in Room 1.L3 (HRI V)
Supernumerary hand for daily/industrial assistance- See Andrea’s presentation on Tuesday, Room 2.R1 (Robot Design)
Fondazione Istituto Italiano di Tecnologia [email protected]
Team HRI2
Adriano Scibilia, MSc studentLeonardo Frezza, MSc student
Fondazione Istituto Italiano di Tecnologia [email protected]
Chapter 3: Feedback InterfacesVisual
M. Lorenzini et al., WeRob 2018, under review
Wansoo Kim
Marta Lorenzini
Fondazione Istituto Italiano di Tecnologia [email protected]
Chapter 3: Feedback InterfacesVibro-Tactile
68.1 mm
54.1 mm
37
.0
mm
A A
17
.3
mm
𝑅 = 123 mm
A - A
BackLower limbUpper limb
Motor controllerMotor
Battery
Microcontroller
W. Kim et al., RAL/IROS 2018, under review
Fondazione Istituto Italiano di Tecnologia [email protected]
Chapter 3: Feedback InterfacesVibro-Tactile
W. Kim et al., RAL/IROS 2018, under review