Genetic Encoding of Robot Metamorphosis: How to Evolve a Glider with a Genetic Regulatory Network
| Autor: | Anne C. van Rossum |
|---|---|
| Rok vydání: | 2010 |
| Předmět: |
Theoretical computer science
Fitness function Computer science Glider Swarm behaviour 0102 computer and information sciences 02 engineering and technology 01 natural sciences Set (abstract data type) 010201 computation theory & mathematics 0202 electrical engineering electronic engineering information engineering Graph (abstract data type) Robot 020201 artificial intelligence & image processing Tuple Reciprocal |
| Zdroj: | Lecture Notes in Computer Science ISBN: 9783642154607 ANTS Conference |
| DOI: | 10.1007/978-3-642-15461-4_62 |
| Popis: | In REPLICATOR [2] powerful reconfigurable robots are designed and constructed. Reconfigurable robots can dock together and form robot organisms. Robot organisms have the ability to morph from snakes into spiders, chairs, swarms, wheels. The problem we are facing is: How to evolve self-organized robot metamorphosis? The metamorphosis graph A = {S,T} is a tuple of S, the set of all possible robot module configurations, and T the set of all transitions between those configurations. A configuration s ∈ S, s = {R,D} consists out of R robots with D connections. If D = ⊘, s denotes a swarm. A fitness function f for reciprocal metamorphosis defines a maximum for a specific cycle in A. A metamorphic fitness function is used that defines maximum fitness for a dynamic body form called a glider: a snake growing its head, and losing its tail a g ∈ A. The evolutionary search process needs to find a self-organized solution for a glider in A. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|