Orientation by Weighted randomness
Résumé
This paper deals with the problem of how motion control can be achieved in very simple systems under a minimal amount of preconditions. The system's design therefore is as minimal as possible, reflecting the basal reflex-arc as observed in biological systems. The model for the movement of the agent is a multiplicatively modified random walk and thus does not represent a diffusion process of Langevin type. The mobile agent shows a reliable and fast homing behavior towards a defined area and finally stays in some defined neighborhood of this area. Moreover, obstacle avoidance is shown to be an immediate result of the system's properties. Keywords: bio-mimetic system's architecture, stochastic motion control, homing behavior, obstacle avoidance