Newsgroups: comp.robotics
Path: brunix!news.Brown.EDU!noc.near.net!uunet!newsflash.concordia.ca!hobbit.ireq.hydro.qc.ca!NetNews.ireq.hydro.qc.ca!gamin
From: gamin@ireq-robot.hydro.qc.ca (Martin Boyer)
Subject: Re: Telerobotics [was: research topics in comp.robotics]
In-Reply-To: noreils@aar.alcatel-alsthom.fr's message of 10 May 93 10:09:35 EDT
Message-ID: <GAMIN.93May12123319@pellan.ireq-robot.hydro.qc.ca>
Lines: 62
Sender: news@ireq.hydro.qc.ca (Netnews Admin)
Organization: Le laboratoire de robotique de l'Institut de recherche
	d'Hydro-Quebec
References: <1993May10.140935.13121@aar.alcatel-alsthom.fr>
Date: Wed, 12 May 1993 16:33:19 GMT

>>>>> Fabrice Noreils writes:

>I am doing some research in Telerobotics for mobile robots also. By
>Telerobotics, I mean the mobile robot and the operator share
>skills and duties and thus the problem is to find (i) the proper
>balance between robot skills and operator skills and (ii) how the
>mobile robot and the operator cooperate in order to achieve a task.

I work on telerobotics for live-line maintenance.  Hydro-Quebec has
successfully demonstrated tasks on 25KV lines with simple
teleoperation, using two Kraft master/slave force/feedback
manipulators.   Here, at the Research Institute, we develop new
technologies that could be used in the field in the next 5 years.

We use a PUMA 760 manipulator, equipped with a force sensor, a laser
range sensor, a trackball, video cameras and lots of software to glue
everything together.  We are getting pretty good results by building a
set of tasks where the operator and the software cooperate in either
shared or traded control.

In shared control, the trackball is used to control some of the
degrees of freedom, while the information from the force and range
sensors drive the other degrees of freedom.  A typical use is
following a conductor in the direction of its main axis; the operator
controls the "axial" speed and the robot is "radially" servoed.

In traded control, on the other hand, precise manipulations are
usually left to the control software, while complex decision-making
processes are done by the operator.  A good example is the closure of
a high-voltage switch, which must be done quickly and precisely, but
not exceeding a maximum force and position.  In this case, the
operator points the laser beam on the switch, and then relinquishes
control to the computer.  A 3D model of the switch is used to position
the robot properly, and impedance control is used to close the switch
itself.  Traded control is an advantage because there is no way the
computer could reliably identify the proper structure in a cluttered
environment, and it would be difficult for the operator to control the
manipulator precisely.


>Of course, there are a lot of areas that need more research in order
>to introduce Telerobotics into industry. The main one remains the
>limitation of the bandwidth of the communication channel.

I disagree.  Telerobotics is also used in areas where communication
delays are irrelevant (nuclear plants, hydro lines) or minimal
(sub-sea).  I believe that proper force-feedback and good contact
stability is more important.

>Finally, I would like to have your comments about simulation and
>Telerobotics. I think it can be interesting to couple a robot
>simulator and a ray tracing system (for visual feedback) in order to
>simulate a real robot.

Electricite de France is doing just that with the help of another
French company, Aleph.

--
Martin Boyer                            mboyer@ireq-robot.hydro.qc.ca
Institut de recherche d'Hydro-Quebec    mboyer@ireq-robot.uucp
Varennes, QC, Canada   J3X 1S1
+1 514 652-8412
