de.mpg.escidoc.pubman.appbase.FacesBean
Deutsch
 
Hilfe Wegweiser Impressum Kontakt Einloggen
  DetailsucheBrowse

Datensatz

DATENSATZ AKTIONENEXPORT

Freigegeben

Bericht

Modeling biological sensorimotor control with genetic algorithms

MPG-Autoren
http://pubman.mpdl.mpg.de/cone/persons/resource/persons83985

Huber,  SA
Department Human Perception, Cognition and Action, Max Planck Institute for Biological Cybernetics, Max Planck Society;

http://pubman.mpdl.mpg.de/cone/persons/resource/persons84072

Mallot,  HA
Department Human Perception, Cognition and Action, Max Planck Institute for Biological Cybernetics, Max Planck Society;

http://pubman.mpdl.mpg.de/cone/persons/resource/persons83839

Bülthoff,  HH
Department Human Perception, Cognition and Action, Max Planck Institute for Biological Cybernetics, Max Planck Society;

Externe Ressourcen
Es sind keine Externen Ressourcen verfügbar
Volltexte (frei zugänglich)
Es sind keine frei zugänglichen Volltexte verfügbar
Ergänzendes Material (frei zugänglich)
Es sind keine frei zugänglichen Ergänzenden Materialien verfügbar
Zitation

Huber, S., Mallot, H., & Bülthoff, H.(1998). Modeling biological sensorimotor control with genetic algorithms (60).


Zitierlink: http://hdl.handle.net/11858/00-001M-0000-0013-E87D-B
Zusammenfassung
Evolutionary optimization of sensorimotor control has lead to matched filter neurons in the visual system of flies that are specialized to certain visual motion patterns. We apply the technique of genetic algorithms in order to model parts of the sensor system and behavior of an artificial agent. The agents are rather simple systems with only four sensors. We will show how genetic algorithms can be applied to evolve simple matched filters that analyze the visual motion information for the task of obstacle avoidance. We compare the agents' sensorimotor control to that of flies. Further we test the optimization performance of the genetic algorithms. We can show that the use of binary or Gray coding has no significant influence on our optimization results and the speed of convergence. Real value coding leads on average to slightly smaller maximal fitness values. The use of a combination of mutation and crossover leads to high fitness individuals and a high fitness population.