Material, architecture and embodied intelligence: an interdisciplinary approach of soft robotics

© ICube

Pierre Renaud, professor at INSA Strasbourg and member of the ICube laboratory (CNRS/University of Strasbourg/INSA Strasbourg), presents the structuring research action (AS1) project entitled ‘Materials, Architecture and Embodied Intelligence’ from the Organic Robotics (O2R) research programme. He explores the interdisciplinary nature of the project and its advances in the field of robotics.

How does AS1 collaborate with researchers who are not specialists in robotics – such as those in anthropology, cognitive science, art, design and philosophy?

Pierre Renaud: The AS1 project was developed by bringing together several disciplines. The issue of flexibility – its benefits and its impact on robot design – fuelled multidisciplinary discussions during the project’s inception and launch. These are very interesting exchanges within the consortium and a strength of O2R today, enabling us to create a community to tackle the subject of robotics together. We are currently managing to develop work across communities that I believe to be original. I am thinking in particular of the work combining art, design and robotics around the development of active foams: robotics, in collaboration with Inria, FEMTO-ST and ICUBE, brings its expertise in modelling, simulation and the use of advanced manufacturing techniques, whilst the arts and design, through ENSAD, contribute insights into the aesthetics, materiality of our devices and the relationship that can be established with a user. This is highly rewarding work for the whole new community we are trying to establish. 

Could you give us an example of a soft robot you’re working on?

P.R.: Devices based on active foams are one example, but it’s difficult to limit myself to just one… Tensegrity structures are another very interesting project in the context of AS1: we use an architecture combining flexible and rigid elements to achieve movement by manipulating the internal forces within the system. This is an approach that remains relatively undeveloped, yet is very rich in potential given its existence in nature, at the scale of a cell or, for example, a bird’s neck, as our colleagues at LS2N in Nantes have studied. In-depth research is therefore possible, to harness the internal forces that pre-stress the robot’s architecture and the way it is controlled. Consequently, a collaborative project between LS2N and ETIS in Cergy has been launched to explore control approaches based on embodied intelligence, aimed at facilitating and improving the control of these relatively novel types of robots.

What characteristics are you looking for in a new material that could be used in the design of a soft robot?

P.R.: In the AS1 project, our challenge is to link the robot’s materials and architecture, and ultimately its control system, in order to master its movements. At the material level, we consider mechanical properties, as is ‘traditionally’ done in the fields of materials science and robotics: stiffness, strength, density. However, we are incorporating other dimensions into the project: appearance and tactile feel, given that these systems are intended to be close to and in contact with humans. This then raises the question of whether these materials can be shaped and used in manufacturing. In this regard, we are currently seeking to take additive manufacturing processes even further—areas in which ICUBE and CEA LIST, in particular, are active—to build the robot or its components by adding material. 

What approaches do you use to tackle the eco-design of robots?

P.R.: We are trying to bring together various areas of expertise and approaches to the issue of design and its environmental impact, as this is an aspect we cannot afford to overlook today. One initial approach is to focus on analysis: assessing what constitutes a robot, going beyond a single metric, such as greenhouse gas emissions. Here, LESC brings an original approach based on deep material science, identifying and highlighting what constitutes a robot, the complexity of its supply chain, the expertise involved, and ultimately the environmental footprint on our planet represented by the production of a product such as a robot. At the same time, we are pursuing an alternative design approach in parallel, going beyond analysis, by choosing to consider the use of bio-based materials, with Inria and LS2N currently working together on issues related to the use of plants.