Towards a real time simulation of Linked Multi-Component Robotic Systems

The autonomous behavior learning in Linked Multicomponent Robotic Systems is an open issue. Single Robot Hose Transport is a limit case of this kind of systems, when one robot moves the tip of the hose to a desired position. The interaction between the passive, flexible hose and the robot introduces highly non-linear effects in the system's dynamics thus making necessary the utilization of accurate geometrical and dynamical models. This paper reports the computational requirements that we have found using these models in a Reinforcement Learning framework and we address a possible solution to mitigate them.