Documento de conferencia
Acceso Abierto

Neuro-Controllers, scalability and adaptation


A Layered Evolution (LE) paradigm based method for the generation of a neuron-controller is developed and verified through simulations and experimentally. It is intended to solve scalability issues in systems with many behavioral modules. Each and every module is a genetically evolved neuro-controller specialized in performing a different task. The main goal is to reach a combination of different basic behavioral elements using different artificial neural-network paradigms concerning mobile robot navigation in an unknown environment. The obtained controller is evaluated over different scenarios in a structured environment, ranging from a detailed simulation model to a real experiment. Finally most important implies are shown through several focuses.

Palabras clave
evolutionary robotics
adaptative systems

Esta obra se publica con la licencia Creative Commons Attribution 4.0 International (BY 4.0)
Imagen en miniatura