Concept & Objectives


Concept and Objectives

This project addresses important and timely objectives outlined by the current ICT call on cognitive systems and robotics (Challenge 2 – ICT 2011.2.1), including the development of artificial cognition  and improving robotic functionality. Completely in-line with the topics of this call, we propose to research and create artificial systems with cognitive capabilities in a framework that is based on engineering and design of robotic systems.

Departing from the traditional robotic manipulation concepts that rely on fixed stiffness distributions, the STIFF-FLOP project takes inspiration from biological “manipulation and actuation” principles as they are, for example, found in the octopus who can turn its links from a completely soft state into a state of precise and, if needed, powerful articulation (see Figure) – an approach that combines advantages associated with both soft and hard systems by selectively controlling the stiffness of various parts of the body depending on the task requirements. Tightly integrating the input from established experts in biology, cognitive sciences, robotics, sensing and medical sciences, this project aims to overcome the drawbacks of current robotic manipulation concepts and to move into a new era for flexible robotics with great promise for many applications areas including minimally invasive surgery. Although we expect this field to bring new challenges when exploring the behaviours of such “stiff-flop” manipulators and how they interact dynamically with their environment, there are many advantages: the ability to squeeze through narrow gaps and openings, inherent compliance leading to increased safety especially when in contact with humans, hyper-redundancy for improved reachability in a obstacle-cluttered environment and increased adaptability to and stability in a possibly unpredictable environment.

With the aim to create an embodiment of an artificial, flexible manipulation device modelled after the octopus limb, the STIFF-FLOP project will develop a new design approach to soft robotic manipulators whose physical morphology and control policies will adapt continuously in order to maintain an optimal symbiosis with uncertain environments and unpredictable task constraints in collaboration with a human operator in applications such as minimally invasive surgery. It is widely argued that cognition is an emerging property of physical and sensory interactions with the environment rather than something that can be programmed. Here we take the view that the uncertainty or stochasticity of the environment itself is an important architect of internal cognition that effectively organizes the body to maintain stable interactions with the real world. We believe that our approach is novel and will lay the foundation for robotic cognitive systems that can creatively face the uncertainty in the real  world by allowing the uncertainty itself to drive the synthesis of cognition that can in turn control the agile body to co-exist with the uncertain world in a meaningful manner.