Dear Aejaz,
I was the colleague who set up the system. The training ended up being over
a period of days as they suffered from chronic fatigue and found it hard to
take the information in. The client had previously used a push button
handset from Possum and it was a case of trying to keep her using something.
Her vision was not functional hence the choice of sicare with voice
feedback; I know there are other alternatives but I like the clarity of
voice on the sicare and the client was able to understand the concept of
voice control. Although it took time she was able to use it effectively in
the end and only had to resort to using the switches occasionally (switches
permanently available). I took advice on training it and did all things that
have been suggested so far apart from using the external microphone, like
recording with various background noise levels and achieved a good
recognition rate. However, as she was able to use a switch with effort (what
Colin Clayton would call an occasional or 1/2 switch) I felt more
comfortable having this back-up in the system.
I am still to be convinced of the reliability of voice recognition for ECS.
I am not suggesting it shouldn't be used but Colin Geggie's approach is good
as you do need people that are aware of the fallibility of the technology
and prepared to accommodate it. I also think that from a service management
point of view, these systems can be more time consuming to install and more
importantly to support. This may be its downfall with respect to other
systems that once installed need less support time, particularly as systems
like SRS can now be interfaced to remotely (yet to find out how effective
this is going to be!). In addition the nature of the system means it is
limited in applicability to client types so with a small number of units per
service may mean stocking additional spare parts and technicians would be
less familiar with units as they would be comparitively less common than the
'standard' units used by the service. Whereas a more versatile unit can be
used across the range of clients a service has and would eliminate these
problems.
I realise these problems are not insurmountable but my personal opinion is
that I have used the system, made it work but feel that they do not offer
enough of an advantage over other systems to make them attractive from a
service point of view.
Gareth Adkins B. Eng. MSc. PGDip IPEM.
Rehabilitation Engineer
Mobility and Rehabilitation Service
Woodend Hospital
Eday Road
Aberdeen AB15 6LS
> -----Original Message-----
> From: Aejaz Zahid [SMTP:[log in to unmask]]
> Sent: 25 November 2002 13:06
> To: [log in to unmask]
> Subject: Re: Voice Recognition for Environmental Control
>
> Hi Jeremy
>
> Just a couple of specific questions about this client:
>
> When your colleague trained the system for this client did they actually
> go at different times/days and therefore perform the training over a long
> period of time? Or, did they do it all in one session by having the client
> try and emulate different situations? Is the backup switch with the client
> at all times and if so, how many times has she used it while she's had
> this system?
>
> Regards,
>
> Aejaz Zahid
> Clinical Scientist
> Department of Medical Physics & Clinical Engineering
> Barnsley DGH NHS Trust, Barnsley S75 2EP
> [T] 01226-730000 x2159
> [F] 01226-208159
>
>
>
> -----Original Message-----
> From: p=NHS NATIONAL
> INT;a=NHS;c=GB;dda:RFC-822=ASSISTECH(a)JISCMAIL.AC.UK;
> Sent: Monday, November 25, 2002 9:09 AM
> To: p=NHS NATIONAL
> INT;a=NHS;c=GB;dda:RFC-822=ASSISTECH(a)JISCMAIL.AC.UK;
> Subject: Re: Voice Recognition for Environmental Control
>
>
> We currently have one client using a Sicare Pilot (we previously had two).
> They were
> both installed and managed by a previous colleague of mine and I have only
> had
> experience of managing the device in recent months, since I have had to
> change its
> configuration.
>
> It is used by an MS patient (middle-aged woman, not technically-minded),
> whose
> speech is now quite slurred and changes through the day. She uses it
> without a
> microphone. She uses it for community alarm, telephone, door entry and
> TV/video. It
> seems to work well, although turning the TV down, when loud, does seem to
> take 2-3
> attempts. The main noise barrier currently is the new washing machine in
> the ajoining
> kitchen, and this may require some new speech learning, with it on in the
> background. I have been pleasantly surprised by just how effective the
> system has
> been and I am now considering supplying it for another MS patient, whome I
> have
> previously discounted it for. The use of multiple patterns, to account for
> different
> intonation/noise etc., seems to work well, although having to access the
> client at
> different times through the day can have practical problems. Obviously any
> such
> system should also have the backup of having a switch input, to at least
> raise an
> alarm.
>
> I have recently been in contact with the Product Manager of Sicare and he
> informs
> me that there is a new version of the software soon to be released,
> although I do not
> know the details of the approach they will be using for speech recognition
> this time
> round.
> Jeremy Linskell CEng,MIPEM,SRCS
> Clinical Engineer
> Co-ordinator, Electronic Assistive Technology Service
> Tayside Orthopaedic and Rehabilitation Technology Centre
> Ninewells Hospital, Dundee, DD1 9SY, Scotland
> tel +1382-496286, fax +1382-496322
> email: [log in to unmask]
> (backup email: [log in to unmask])
|