Evolving-Controllers Versus Learning-Controllers for Morphologically Evolvable Robots

Autor: Miras, Karine, De Carlo, Matteo, Akhatou, Sayfeddine, Eiben, A. E., Castillo, Pedro A., Jiménez Laredo, Juan Luis, Fernández de Vega, Francisco
Přispěvatelé: Artificial intelligence, Network Institute, Artificial Intelligence (section level), Computational Intelligence, Knowledge Representation and Reasoning, Castillo, Pedro A., Jiménez Laredo, Juan Luis, Fernández de Vega, Francisco
Jazyk: angličtina
Rok vydání: 2020
Předmět:
Zdroj: Applications of Evolutionary Computation: 23rd European Conference, EvoApplications 2020, Held as Part of EvoStar 2020, Seville, Spain, April 15–17, 2020, Proceedings, 86-99
STARTPAGE=86;ENDPAGE=99;TITLE=Applications of Evolutionary Computation
Applications of Evolutionary Computation ISBN: 9783030437213
EvoApplications
Miras, K, De Carlo, M, Akhatou, S & Eiben, A E 2020, Evolving-Controllers Versus Learning-Controllers for Morphologically Evolvable Robots . in P A Castillo, J L Jiménez Laredo & F Fernández de Vega (eds), Applications of Evolutionary Computation : 23rd European Conference, EvoApplications 2020, Held as Part of EvoStar 2020, Seville, Spain, April 15–17, 2020, Proceedings . Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12104 LNCS, Springer, pp. 86-99, 23rd European Conference on Applications of Evolutionary Computation, EvoApplications 2020, held as part of EvoStar 2020, Seville, Spain, 15/04/20 . https://doi.org/10.1007/978-3-030-43722-0_6
DOI: 10.1007/978-3-030-43722-0_6
Popis: We investigate an evolutionary robot system where (simulated) modular robots can reproduce and create robot children that inherit the parents’ morphologies by crossover and mutation. Within this system we compare two approaches to creating good controllers, i.e., evolution only and evolution plus learning. In the first one the controller of a robot child is inherited, so that it is produced by applying crossover and mutation to the controllers of its parents. In the second one the controller of the child is also inherited, but additionally, it is enhanced by a learning method. The experiments show that the learning approach does not only lead to different fitness levels, but also to different (bigger) robots. This constitutes a quantitative demonstration that changes in brains, i.e., controllers, can induce changes in the bodies, i.e., morphologies.
Databáze: OpenAIRE