Abstract
(Englisch)
|
Despite their well-known advantages, traditional engineering methods are inflexible and unreliable when compared to biological organisms. We propose a novel architecture for the development of artefacts that will be inspired by cell-biology and based on simple building blocks that allow for behavioural and morphological co-adaptation. Our architecture will allow transitions from simple accumulations of similar building-blocks to self-organized agents made of task-specialized building-blocks. We will exemplify the potential of the new design standard through the construction of robotic systems that can change morphology. Simulations will be used to extract and transfer the required biological mechanisms to the physical implementation. The implication of our novel architecture will be a design method for the construction of artefacts with a magnitude of flexibility and robustness far beyond any feasible future perspective of conventional engineering and a great variety of applications.
Objectives:
This project concerns the development of a fundamentally new design standard, which will allow robust and efficient morphological development of artefacts. In order to allow end-users of our approach to design new artefacts in an easy manner, we will provide an open standard architecture made up of simple building blocks, simple connections, and simple interactions. Systems comprising hundreds of basic building blocks exhibit self-assembly, self-repair, and differentiation based on location. The Hydra standard will provide end-users with a straightforward technique to exploit holistic behavioural and morphological adaptation in the design of new artefacts. The project will exploit biological insight into the mechanisms of biological self-organisation, of morphological transitions between different life stages of organisms, and the concepts of self-replication, self-repair, and mutagenesis in the search for parsimonious choices of components and versatile interactions between them.
Work description:
Inspired by developmental biology, the project will create a physical model of cellular self-assembling organisms in 3D space, which will co-evolve behaviour, morphology, and perception in response to their environment. Each cell component will cooperate with other cells and will be able to make a choice regarding its investment in abilities such as movement, perception, and communication. It is expected that this design will ultimately lead to task specialisation of the cells (division of labour) due to temporal and spatial constraints, such as queuing and bottleneck effects as known from studies on social insects. The resulting simulated agents are then to be translated into the real world, which requires the availability of a highly modular 'rapid prototyping robot building kit'. In order to enable the exploitation of morphological change in addition to neural plasticity, the components will have to allow for autonomous self-reconfiguration. In order to achieve the latter, principles of self-organisation will be studied and employed. Initially, the building blocks of this robot building kit will be coarse-grained, on the order of centimetres. After a research cycle of building-block prototyping, the project will produce several hundreds of standardised building-blocks for experimentation and a demonstration of the new Hydra architecture.
|
