The pick of the crop of YouTube videos featuring Aldeberan Robotics’ NAO Robot
Aldeberan’s promotional video featuring NAO interacting with a person in various ways. Most likely none of these behaviours are fully formed or work exactly as shown. For instance it’s doubtful NAO can recognize his friend’s face from across a room in real life. Of particular interest is the section (starting at 2 minutes) where NAO walks, showing the original and the new optimized walking algorithms in one shot.
A demo of NAO by Aldeberan’s founder Bruno Masonnier. It’s his personal enthusiasm for robotics and the care they have taken in crafting NAO that makes me enthusiastic about the possibilities of developing for this platform. The interactions between Bruno and NAO are fairly natural at points. In particular watch NAO frustrated after Bruno tells him to be quiet.
NAO using the prototype self-charging station. Watch the disconnect behavior near the end for a treat.
Two NAOs interacting. A good video showing NAO grasping an object on the ground, walking, etc. and using a NAOcode marker on the trash can to locate it. Again some of this interacting is a bit “staged” (for instance can one NAO understand the other via voice alone?) but it’s a good indication of what NAO should be able to accomplish when fully programmed.
NAO executing the walk with me behaviour. Feedback from the arm is used to plan NAO’s path forward, back, and side-to-side.
NAO taking and handing back an object. While this looks trivial, notice the subtle interactions, checks, and balances that have been built into the process so both the robot and the human know what’s going on. This is the 5th version of this particular behavior. The older one’s are on youtube as well for comparison
NAO climbing a ladder. They admit that it’s a behavior built for this particular ladder (for instance different rung spacing might make it fail) but you can see that the robot is physically capable of lifting it’s own weight, grasping the rungs, and climbing.
An ambitious development team creating a behavior to allow NAO to play the children’s game Connect 4. See their channel for examples of NAO actually playing parts of the game. This one is a work in progress and the video represents where they want to end up. It also highlights how many complicated steps there are to things that we humans take for granted. Note also that NAOs fingers are held together with a small rubber band to allow him to hold the Connect 4 pieces.
NAO using the 8 force sensing resistors in the bottoms of his feet to react to sideways pushing and to balance on tilting surface
Normally NAO’s joints are set to full stiffness (a setting of 1) when he executes movements. This gives the greatest control, but also uses the most power and overheats his servos the quickest. With NAO’s stiffness set to 0 he is normally limp and can be posed by hand. In this inspiring video they are apparently positioning NAO to walk then setting his stiffness to zero right as the walk cycle begins, then returning it to 1 again at the end of the animation. Check the poster’s youtube channel for more experiments of NAO walking with various stiffness settings
And of course, NAOs playing soccer. The University of Bremen team write their own walking, kicking, etc algorithms from scratch, replacing the built-in walk cycle NAO normally uses. Especially impressive this the rearward kick executed by sliding NAO’s foot over the ball in reverse.
NAO asks for help getting up. Apparently they are sensing the arm movement to initiate the standing action. Also on the second attempt, they lift just one of NAO’s arms and he responds differently.