Skip to Main Content
In this paper the importance of the feedback from sensors in the neuro-controller of the N.E.Me.Sys. robot is proved. N.E.Me.Sys. is a legged rover for planetary exploration fully controlled by artificial continuous time recurrent neural networks (CTRNNs) designed with evolutionary algorithms. The most interesting result of this work is that it explain how a leg equipped with contact sensors is able to better cope with all the unevenness of the terrain. Moreover it is possible to verify that a higher influence of the sensors in the generation of the internal dynamics of the neural controller increases the adaptability and robustness degrees versus uncertainties and off-design conditions.