This project aims at developing robots that can co-exist and co-operate with people, enabling a physical human-robot interaction which is dependable and safe: in a word, to make robots and humans PHRIENDS.
The next generation of robots will interact with people directly. Human-Robot Interaction (HRI) will certainly happen at the cognitive level (cHRI), fundamentally concerning communication between human and robot through the many channels available to us (video displays, sounds, watching each other’s motions, spoken language, or even gaze direction or facial expression). However, robots are distinct from computers in that they physically embody the link between perception and action. Hence, the most revolutionary and challenging feature of the next generation of robots will be physical Human Robot Interaction (pHRI). In pHRI, humans and robots share the same workspace, come in touch with each other, exchange forces, and cooperate in doing actions on the environment. Robots designed to cooperate with humans, such as e.g. in assisted manipulation, collaborative assembly, domestic, entertainment, rehabilitation or medical applications, must fulfill different requirements from those typically met in conventional industrial applications. Typical conventional robot systems and applications require fast motions and absolute accuracy in positioning and path following and avoid using expensive and error-prone sensors at the tradeoff of having carefully designed and modeled environments. In sharp contrast, the before-mentioned future application areas could use (extra) internal and external sensor systems and could trade in certain performance characteristics of the robot system to drastically increase dependability and overall safety in unpredictable dynamically changing environments. Given such a discrepancy in requirements this project aims at substantially advancing the state of the art in robot design and control to enable applications that require intrinsically safe physical human-robot interaction.

PHRIENDS is about developing key components of the next generation of robots, which have to meet the strictest safety standards, yet also to deliver useful performance. This poses new challenges to the design of all components of the robot, including mechanics, control, planning algorithms and supervision systems. The approach advocated by this project consists in an integrated co-design of robotic systems for safe pHRI, which revolutionizes the classical design paradigm of industrial robots – rigid design for accuracy, active control for safety – into a new one: design robots that are intrinsically safe, and control them to deliver performance. Although the scope of this project cannot encompass the integration of complete robot systems, PHRIENDS will achieve fundamental advancement in the direction of safe pHRI, for it will create and deliver:
A) new actuator concepts and prototypes;
B) new dependable algorithms for supervision and planning;
C) new control algorithms for handling safe human-robot physical interaction and for fault tolerant behaviour.

Furthermore, PHRIENDS will
D) integrate these components in functionally meaningful subsystems;
E) evaluate quantitatively components and subsystems via experimental testing;
F) contribute to the ongoing effort of international bodies towards the establishment of new standards for collaborative human-robot operation.

These ambitious objectives will be achieved through the presence in the consortium of three academic groups, two research laboratories, and one industrial partner who have specific competence in human-robot interaction and a substantial history of collaboration in the field. This will enable us to fundamentally advance the state-of-the-art in two complementary directions:
1. integrate new algorithms in existing manipulators, produced by our industrial partner, and allow new paradigms for pHRI in service and industrial environments;
2. design, implement, test, and optimize the core components of the next-generation, intrinsically safer robot arms.
Eventually, all project activities will culminate in experimental platforms and test-beds. The first leg of the project will lead to an experimental platform that is based on the new KUKA light-weight robot arm. Its controller will integrate in a prototypical fashion the newly developed algorithms. A demanding pHRI application scenario will be set up to demonstrate its applicability. The second leg of the project will lead to test-beds that will be used to evaluate and optimize safety and performance characteristics of a new generation of intrinsically safe robot arms.