Evolution of Controllers for Robot-Plant Bio-Hybdrids: A Simple Case Study Using a Model of Plant Growth and Motion

In evolutionary robotics methods of evolutionary computation are applied to evolve robot controllers [1]. Evolutionary robotics is also our method of choice in the EU-funded project flora robotica [6, 8], which pushes research towards the evolution of a broad variety of artifacts and contraptions [4]. We are investigating how a distributed robot system and a group of biological plants can be tightly coupled to generate synergies and Ąnally result in a bio-hybrid system. We want to create a co-dependent and self-organized system with closely linked symbiotic relationships where plants support robots, for example, by providing scafolding and robots that direct plant growth towards desired areas. In addition to plant growth there is also plant motion which is often ignored, probably due to the fact that the motion of plants is slow compared to that of animals. It can be diicult sometimes to distinguish between a plantŠs motion and growth as both happen concurrently and usually in similar ways. They can difer, however, fundamentally in terms of time scales. Another important diference is the fact that motion can be reverted while growth is mostly permanent. Besides plant growth we also want to harness plant motion in our robotplant systems.