Roboskin

Rationale

Tactile sensing is strategic for safe interaction of robots with humans, objects, and possibly in unstructured conditions. As a matter of fact, contact sensing provides the most important and direct feedback to control contact both in case of voluntary and non-voluntary interactions with the environment.

Beyond the classical robot interaction tasks (e.g., peg-in-hole problems) where interaction is expected or planned at a specific robot location (typically at the end-effector tip), more advanced applications require more complex forms of interaction (e.g., whole hand or whole arm grasping and manipulation, gait stability control etc.), where the location and the characteristics of the contact could not be exhaustively predicted or possibly modeled in advance. Furthermore, interaction control can have a more complex perceptive role as for instance in the case of programming by demonstration or skin based social cognition tasks, where the modes of interaction arise from the human-robot contact, and convey the information related to the robot task.

In order to tackle these new issues both at control and perceptive levels, sensor systems have to be developed to measure interaction phenomena over large contact areas. Furthermore, these sensor systems have to be properly interfaced to motor control modules to ensure reactive and safe interaction of the robot with the environment. Finally, appropriate robot perception and cognitive strategies must be implemented based on the underlying sensing structure, in order to enable meaningful human-robot interaction tasks.

The objective of RoboSKIN is to improve the ability of robots to act efficiently and safely during tasks involving human-robot interaction. To this aim the project focuses on:

1. The study of sensing technologies and methodologies for the development of distributed and modular components for building robot skin.
2. The study of control and perception tasks required to develop cognitive mechanisms that use tactile feedback to improve human-robot interaction capabilities.

Therefore, this project aims at improving the robot capabilities to act efficiently and safely in unconstrained environments as well as their ability to communicate and interact with each other and with humans through touch and contact.

At the end of the project we wish to prove that:

1. There exists a fundamental role of tactile sensing and perception, for the implementation of a class of application relevant human-robot interaction tasks, which motivates the development of large scale robot skin systems.
2. It is possible to implement procedures and methods for the development of large scale robot skin systems, using state of the art technologies and accessible servicing facilities, which can be tailored for different robotic platforms.
3. It is possible to develop a middleware bridging the tactile hardware with the perceptive and the control modules, also supporting the implementation of high level skin based cognitive robot interaction tasks.
In order to meet these goals the project pursues three intertwined objectives addressing different engineering and control problems. At the lowest level, the development of the robot skin technology will require mostly robotics, electronics, mechanical and systems engineering. At the middle level the implementation of algorithms for tactile data processing, reactive skin based control, and perception; learning of interaction phenomena will involve expertise in the areas of control and software engineering as well as robotics. Finally at the highest level, where specific applications will come into play involving specific high level tactile based skills, expertise in the areas of applied robotics, computer science and engineering will be required.