Skip to Main Content
This paper proposes a parent-children type robot system that moves in unstructured environments. The parent robot works as a leader of the system. The children robots work as sensors to sense their environments while touching them. The parent collects the sensory information of its environment and generates a map. In order to express the map effectively, this paper applies a structured neural network to memory. The neural network learns the sensory information incrementally. While using the network, the parent robot determines their behavior. On the other hand, the children are disposable. When some of the children malfunction because of their dangerous environment, the remaining children compensate for them and continue to work. Simulations are performed to show the effectiveness of the proposed system.
From Perception to Action Conference, 1994., Proceedings
Date of Conference: 7-9 Sept. 1994