# help with a #SENSOR# problem.. any info would be great, thank you.

N

#### Nick

Jan 1, 1970
0
Hi guys,

I know this is an engineering ng but I was hoping you could still provide
some
help as your area of expertise may possibly relate to what I'm after.

I'm a 4th year interactive design student working on a project which
involves the use of sensor triggers. The triggers are used to signal the
interactivity of the target and a device (intelligent lighting, audio,
projections, and environmental control systems) but we are having huge
problems finding potential sensors that can do what we want. Ideally, we
would like to track individuals or groups through a space 90m long, 9m wide,
and 10m high (essentially a tunnel) and interact with these bodies as they
move through the space. It is a public space so the sensors need to be
mounted on the walls high enough to be out of reach or encased in a unit to
protect them.

At first we looked at simple infrared beams like most retail stores use but
this would not give us the resolution required in the interactivity
elements. We have considered mounting sensors on the roof pointing down
but this would require far to many as the space is so large. We have
considered
also vision tracking but we do not know much about it as the technology is
in its
infancy but if anyone is familiar I would greatly appreciate a reply.

We ultimately want to achieve a system in which we can return to a
server system a coord (x,y) value of every body in the space. We were also
looking at using basic sensor mats on the floor but the client is not happy
with this and prefers the sensors to be largely no visible. So I was hoping
some sort
of proximity sensor that can tell where a target is in relation to itself
and then by using
a combination of software and triangulation techniques identify where it is
in the space
and thus return coord of that object. Am I totally left field here in trying
to find something
that will do this???

If anyone has any suggestions on how we can achieve this or come close even
I very much would like to hear your opinions. The budget is not really a
concern
at this point as the clients are willing to spend a fair amount but of
course there will
be a limit but fortunately they do realize what they are asking and
consequently have already considered the potential for an expensive costing
to develop this tracking system (a very rough estimate is around AUD$8K for just the sensor equipment; this can be modified but by how much I'm not sure as yet). Any help would be greatly appreciated. Thank you. Thank you. Kind regards, Nick. N #### Name Goes Here Jan 1, 1970 0 Good Luck. That is a tough problem. Mats are the real low cost solution. Like they use on Busses to detect people stepping on them, IR is next bet. RF is no good. Lazer or optic is too expensive. Cant control reflective of persons, unless you tag them with a tape. Assuming your needed resolution is about 2 by 2 ft. (one person) N #### Nick Jan 1, 1970 0 Hi, Yeah I think IR will be the best. Do you have any further information? Nick. A #### Anthony Jan 1, 1970 0 Check out Cognex. You are looking at some initial cost (about$6k per unit), but it can do
the recognition AND the measuring. It programs in what is basically an
excel spreadsheet, and interfaceable via ethernet. You can also have
discrete I/O if so desired.
I do not consider vision systems to be in infancy...we are using them all
over the manufacturing floor for inspection, detection and measuring. They
are reliable and work well.

--
Anthony

You can't 'idiot proof' anything....every time you try, they just make
better idiots.

Remove sp to reply via email

Replies
1
Views
3K
Replies
0
Views
363
Replies
5
Views
933
Replies
1
Views
652
Replies
1
Views
828