Romeo is Aldebaran's second robot. Much taller than NAO, Romeo is destined to help elderly people, and he's being developped in our offices. We gave the opportunity to our forum members to ask questions about Romeo to Rodolphe Gélin, head of collaborative projects at Aldebaran, and here are his answers :
Is Romeo going to be strictly aimed for the care of the elderly or do you see it targeted as a common household servant?
Romeo is partly funded by collaborative projects dedicated to assistance to elderly people. So we are putting a special focus on assistance for elderly people : memory, locomotion, safety but also entertainment. And because we believe in the concept of “design for all” (what is useful for some should be useful for everyone), we keep in mind that elderly people will not be the only users of our robots. But as for Nao, we don’t want to see Romeo mainly as a servant but as a companion that cares of its human friends. Of course, the size of Romeo will offer new capabilities to it (see next question).
What physical capabilities will Romeo have? Will he be able to open close doors, move and lift objects of a few kilograms?
The size of Romeo has been designed to make it tall enough to be able to grasp objects on a table and small enough to look a seated person without being impressive. We wanted to have it able to walk at a reasonable speed to accompany a person and able to climb stairs. Its hand is designed to grasp everyday objects (bottle of water, cans, door handle…). Each arm has been dimensioned to carry a 1,5L bootle. So the robot should be able to open doors and to carry objects up to 3kg with two hands.
How did the experience with Nao helped for Romeo ?
The engineers who designed Romeo used to work (or keep on working) on Nao. Their expertise was very useful in the development of Romeo. Some parts of development have been made collaboratively for Nao and Romeo, for instance the ATOM CPU (replacing the previous Geode processor) or the HAL (hardware abstraction layer). Some joints of Romeo (neck, elbow, shoulder) are inspired by the cylindro-spherical joint of Nao (but bigger). And the joint that moves the head of Romeo(at the top of the neck), is an ankle of Nao ! But of course some parts of Romeo are completely new : the legs actuation, the vertebral column, the moving eyes.
How close is the API software for Romeo compared with Nao? Is there a RomeoQi?
The development environment of Romeo is exactly the same as the Nao’s one. Based on NaoQi, you have Choregraphe. From the developer point of view, it should not change anything even if, behind, they are some special functions to deal the fact that the software can run on 4 separate processors.
What is the most challenging part in Romeo conception (sofware/hardware, security/robustness/AI/etc), compared with Nao ?
From a technical point of view, the design of the legs was the most challenging subject. The brand new actuation system, that we adopted from the French research Institute CEA, requires very accurate manufacturing. We have chosen to integrate it in an exo-skeleton structure made of carbon fiber. We learnt a lot about fabrication of carbon fiber structures that is quite delicate. And the choice of the exo-skeleton architecture made assembly and debugging of the leg very time consuming. The torque control that we decided to implement in each control board of each joint was a challenge for our electronics and firmware teams. This torque control is a major issue concerning the safety of the robot. But everything we learnt will be very useful on our next robots.