Etude sur l'intérêt des modèles biologiques de réseaux de neurones pour la synthèse de rythmes locomoteurs adaptatifs.

摘要: This article presents a preliminary study for work that aims at increasing the capacity of adaptation legged robots. The controller consists one neural networks, synthesised by evolutionary algorithm, to which adaptive mechanisms inspired biology are implemented. We compare performances controllers according three models neurons with biological complexity. thus show interest dynamic parameter must be calculated from rather than advance. incorporate then these in robot made up leg 3 dof pushing body is sliding on rail. allow line and cycle locomotion maintain performance speed robot.