Project

Back to overview

Embryogenic Evolution: From Simulations to Robotic Applications

English title Embryogenic Evolution: From Simulations to Robotic Applications
Applicant Eggenberger Hotz Peter
Number 118127
Funding scheme Project funding (Div. I-III)
Research institution Institut für Informatik Universität Zürich
Institution of higher education University of Zurich - ZH
Main discipline Information Technology
Start/End 01.10.2007 - 31.01.2010
Approved amount 95'450.00
Show all

Keywords (13)

embryogenic evolution; developmental biology; evolution; simulation transfer to the real world; morphogenetic processes; morphology; cognition; genetic regulatory networks; ontogenetic and phylogenetic development; self-assembly; self-repair; adaptivity; robustness

Lay Summary (English)

Lead
Lay summary
Our ultimate scientific goal is to understand the necessary evolutionary and developmental concepts required to generate complete agents. The promising results, illustrated in the successfully completed Hydra project, are now ready to be fully explored on the following platforms: with the self-assembling Hydra modules, we plan to generate complete agents by evolving the genome in each physical building block. Since the artificial cells are ahle to mimic natural morphogenetic processes, as shown in the Hydra project, we would now like to implement these results on thereal-world building blocks: the terrestrial atrons and the aquatic hydrons. In natural evolution generations of species' morphologies constantly change due to mutation. Since the species' cellular substrate (neurons, sensory tissue, effectors) is capable 6f learning, the agent can cope with these morphological changes. This is an advantage for evolution, allowing for quicker convergence due to the morphological and neurological independence. Evolution does not evolve specific, hard-wired tasks hut adaptive solutioJlS capable of dealing with different morphologies. By mimicking this principle we can evolve neural controllers in simulation and easily transfer these to the real world, because growing, learning and therefore adaptive cognitive systems are evolved. To this end we plan to evolve such cognitive systems for morphologies given by the different platforms we intend to use, such as pneumatic manipulators exhibiting non-linear dynamics and a conventional industrial robot.
Importance Should the EES prove successful on the above mentioned platforms, a contribution toward the autonomous control of robots could be achieved, giving the opportunity to further understand the underlying principles of evolution.
Moreover, the project is highly interdisciplinary, involving biologists, roboticists, computer scientists and control engineers. Given the success of this novel methodology, underlying morphogenetic processes could be better understood and new technologies could be created.
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Associated projects

Number Title Start Funding scheme
109864 Embryogenic Evolution: From Simulations to Robotic Applications 01.10.2005 Project funding (Div. I-III)

-