Newsgroups: comp.robotics
Path: brunix!uunet!stanford.edu!Csli!colleen
From: colleen@Csli.Stanford.EDU (Colleen Crangle)
Subject: Re: Verbal interaction with robots
Message-ID: <1992Feb26.220310.20170@Csli.Stanford.EDU>
Keywords: speech, natural language communication
Organization: Stanford University CSLI
References: <1992Feb25.012201.22744@Csli.Stanford.EDU> <1992Feb26.003051.12009@uokmax.ecn.uoknor.edu>
Date: Wed, 26 Feb 1992 22:03:10 GMT
Lines: 19

In article <1992Feb26.003051.12009@uokmax.ecn.uoknor.edu> rwmurphr@uokmax.ecn.uoknor.edu (Robert W Murphree) writes:
>
>I talked to a handicapped engineer at the Palo Alto VA about voice
>interfaces with mobile wheelchair -he warned me that the verbal  
>bandwidth was too slow for real time control of a wheelchair, also
>the mistaken words could result in disaster-dumping the patient.

Verbal control of a wheelchair is really not a good idea, in part
because ordinary language doesn't have a good selection of "navigation"
terms, not enough for the kind of fine control you'd need to
navigate competently.  Trying to use specially designed artificial
languages has proved frustrating to users.  Controlling a mobile
vehicle, not a wheelchair, with natural-language commands appears
to be different, however.  There we tend to direct the vehicle as we 
would another person, by reference to objects in the environment, with
some added cautions and corrections such as 'Don't hit the lamp' and
'Slow down as you approach the doorway.'
 

