Research activities of the HRI2 laboratory can be divided into two broad categories:
- Modelling and analysis of the human physical interaction behaviour
- Human in the loop robot planning and control
While the first line of research investigates for reliable and intuitive human-robot interfaces relying on or inspired by the human motor functionalities, the latter aims at the achievement of the enhanced human-robot-environment physical interaction performances.
I. Human-Robot Interfaces
Learning Human Sensorimotor Control for Advanced Human-to-Robot Skill Transfer
The research plan in modelling and analysis of the human motor behaviour is to capitalize on the group's expertise in the development of reduced complexity models that can be used for real-time control of the robots or assistive devices in environments with dynamic uncertainties. In this direction, the use of bio-signals (e.g. EEG and EMG), are explored as control inputs for the estimation of the human motor control parameters and intention for robot control. The research activity in this area is divided into two main categories: Brain-machine and Body-Machine Interfaces.
Stability and Robustness of Human-Robot Interfaces
Research activities in this field will investigate the stability and robustness of the developed human-robot interfaces to ensure safety and reliability. This consideration becomes of high importance when the human comes in a direct contact with the robot or assistive device, or co-exists and cooperates with it to achieve a certain task.
II. Human-in-the-loop Planning and Control of Robotic/Assistive Systems
Robot Planning and Control
The underlying human motor control concepts are integrated into the target robots’ control architectures. This potentially generates human like interaction performance and redundancy resolution, while decreasing the robot control burden.
By utilizing the expertise and achievements mentioned in I and II, the lab research activity will target several applications that involve and require enhanced human-robot interaction including:
In this field, the main research activity of the lab is to extend the teleimpedance control concept to larger degrees of freedom: from hand-arm and dual-arm systems to whole-body humanoid implementation.
Rehabilitation and Power Augmentation
The research activity in this area will explore the use of bio-signals such as surface and intramuscular EMGs or for accurate control of the prosthetic/power augmentation devices. In this direction, a shared level of autonomy between the user intention and device controller will be implemented that will potentially result in a desired interaction and manipulation performance. The use of reduced complexity modes (e.g. the concept of motor synergies as explained above) in intuitive control of the prosthetic devices is being explored.
Cooperative and Collaborative Human-Machine Systems
A key factor in the realization of intuitive human-robot cooperation is the ability to understand and estimate both at the motion and impedance regulation level, the human intention in the robot control framework. Starting from the fundamental human-robot collaboration paradigm, the group will extend this to the case of human-humanoid cooperation and apply this to the several humanoid platforms of IIT including COMAN, iCub, and WALK-MAN.
HRI2 laboratory establishes a close collaboration with
- Humanoids and Human Cantered Mechatronics Lab. (N. Tsagarakis),
- Soft Robotics for Human Cooperation and Rehabilitation Lab. (A. Bicchi),
- And several other labs from the ADVR and the iCub facility of the IIT.