Mo re in de ta il s, th e in te rf ac e is up da te d wi th ic on s corresponding to the number and types of objects recognized by the robot perception module and the decision tree used to allow the user to generate the messages for the robot is also updated with the possible actions to perform on the base of the different kinds of recognized objects. In add iti on, a gra phi cal use r inter face (GUI ), develo ped withi n the BCI2 000 fram ework, all ows the use r to sel ect the obj ect s to man ipu lat e and the ac ti on to pe rf or m suc h GU I ha s bee n in te gr at ed wi th th e ro bo t perception module to dynamically update the possible actions to perform on the base of the information received at run-time. Controlling such a redundant system with a BCI by using the proposed novel approach would be, indeed, an overwhelming operation for a user, thus the system has to own a certain level of aut ono mo us cap abi lit ies. The main contribution of the pap er is the int egr ati on bet we en the P30 0 par adi gm and a control architecture that allows to control a system with a large number of DOFs, autonomously taking into account low-level sa fe ty ta sk s bo th at jo in t an d op er at io na l spa ce le ve ls. C control strategies to handle the whole-body motion.
0 Comments
Leave a Reply. |