|Sep26-07, 10:45 PM||#1|
I had this idea a while ago but having just finished reading an article about the recent increase in college-aged entrepreneurs, I was inspired. At present, I am a freshman math major at a small state college in Vermont. I plan to transfer into a competitive Economics/Mechanical Engineering program next year but for now, my technical knowledge is limited to say the least.
The Reason for my Idea:
Any organization that is responsible for maintaining a large inventory needs a system to keep track of this inventory. With RFID and bar code scanning, it fairly easy to track an item when it either leaves or arrives at a facility. However, when the item is on-site, tracking can be a little more difficult. Although I'm sure that many business facilities are more efficient than libraries, the library will serve as a good example. The Dewey-Decimal system certainly works, but there is an easier way.
1. You walk into the library and go to one of the computer catalogs
2. You search for the book/books you want and click an on-screen button "Locate Now"
3. The computer displays a message "Row 300 Look For 2 Short Blue Flashes"
4. You make your way to Row 300 and a find section of one of the LED strips running along the edge of each shelf that is blinking with two short blue flashes
5. Directly above the flashes is the book you are looking for
The same concept can be applied to a warehouse, video/music store, large physical filing system, etc.
How it Might Work:
-RFID tags on the item relay the location to nearby receivers (would you need 3 to triangulate or is there another way?)
-A computer system takes the location and calculates the distance from the starting point of the LED strip to the item and which LED strip it will use
-Different combinations of LED colors and flash timings could differentiate between different users/requests. For example: Command= "Locate all items received on March 14 2007" Items that need to be shipped flash red and items that don't need to be shipped flash blue
The kind of LED strip I'm thinking about would look something like this although it would need to be more sophisticated with multiple colors and individual wiring not parallel. Again, I know very little about EE but that doesn't sound too difficult. The hard part would be setting up the system that determines the location of the item and then writing the software that takes that information and makes us of it.
If anyone sees any potential in this idea, please reply. Otherwise, I will be found back at the drawing board. I'm sure that I will post more ideas in the future. You have been warned All feedback appreciated. Thank you.
|Sep27-07, 06:30 AM||#2|
I don't know a whole lot about RF tags and such, but the idea is sound.
From a technical standpoint, how would the system handle a situation where a product is misplaced? Would the RF tag take care of that?
From a business viewpoint, I would think that this would be a difficult thing to retrofit into an existing location. Have you given any thought to that aspect or were you thinking along the lines of new installations where you could easily put it in as a design feature?
|Sep27-07, 07:12 AM||#3|
The system could also (I suppose) be programed to ID misplaced books/products with a specific flash/color sequence so that the librarian/inventory manager could easily locate them and put them back in their appropriate place.
I also do not know a whole lot about the limitations of current RF technology but I am hoping that it is possible to have a series of cheap short-range receivers installed in the shelf that are capable of identifying multiple RFID tag numbers. For example if there were 2 receivers per meter, each receiver would need to be capable of registering say, 10-20 different book RFIDs and their relative positions (I guess that within 1/2 meter would be acceptable but it would be preferable to narrow that down to 5-10 cm.)
The receiver would register RFID tag xxxxx059 and then the system would register xxxxx059 on receiver 0094 which it knows is located on row 300, 2nd floor. If xxxxx059 popped up on a different receiver, the system would automatically recognize that it is out of place and do something to correct the error i.e. call for the librarian (aren't we supposed to control the machines and not the other way around?)
However, how difficult would it be to install new shelves? Couldn't we just close a section of the library for 1-2 days, unload all of the books while simultaneously giving them their new RF tags and entering them into the system (okay maybe more like 5-7 days), install the new shelves, wiring and mainframe and then replace the books/items one section at a time so as not to handicap the whole facility?
I think that the best way to do it would probably be to do as you said and install it right along with a new building.
Thank you for your response.
|Sep27-07, 10:43 AM||#5|
Welcome to the PF, cnichols. Keep on thinking up new ideas, and keep on working out the details involved, and prototyping parts of the ideas, and learning and improving on your ideas. Keep pushing down that path, and you will definitely do well with some of your ideas.
This present idea has some promise in some forms, but I think there may be too many practical problems with just a straight location finder using RFID or other RF technologies in close quarters like in a library. The biggest issue, I think, is multipath. The frequencies that would be candidates for this system will have multipath issues (reflections from nearby metal structures like bookshelves, metal wiring in walls, etc., add and subtract from the received signal at the receivers, confusing the distance calculations). With enough spatial diversity (multiple antennas) at the receivers, you might be able to figure out the math to get reasonable position data, but you will have to work through the math to figure out if that is realistic or not.
A more workable system might be to more actively track books as they move about the library, as you pick them up with proximity RF sensors as they move through doorways, onto and off of carts, and onto particular shelves. You could also have portable proximity detector handheld units, that patrons could carry with them if they were looking for a book in a room, but were not able to find it where they expected it.
Another variation (a little weirder....) would be to have several cooperating directional antennas that scanned actively with narrow beams to ping for books, and maybe even used the coincidence of both (or 3) beams to activate RFID tags for return pings (that would be picked up either with the beam antennas, or with a simple omni antenna that was separate. For the steerable activation beams, you could either use Yagi antennas inside enclosures (so the antennas waving around wouldn't be distracting to the patrons), or if you got real fancy (expensive?), you could use phased array antennas for the activation beams.....
Hmmm. That last solution might work for the person posting a similar question in the other thread, for locating RFID tags in a room. Think I'll go back and post something there about this....
|Sep27-07, 12:21 PM||#6|
I'd be interested to know if there are antenna systems that can be installed under carpeting and to what extent antenna systems can detect proximity. As a practical matter the ability to retrofit old facilities would be an important consideration in devising tagging systems.
|Sep27-07, 12:36 PM||#7|
Thanks for the welcome berkeman
hmmm getting warmer check this out http://www.merl.com/projects/rfig/
Berkeman, your first couple ideas are interesting but not exactly what I'm looking for. I have the feeling that although there would undoubtedly be some people like those of us who spend our days with noses buried in textbooks who would find it really 'neat' to have a hand held gizmo to find a library book but for most people, it would be a nuisance.
However it seems that this approach is exactly what the people at Mitsubishi (above) are pursuing albeit for a different application.
If we could figure out a way to incorporate this optical sensing/RFID system into a stationary projector and have a computer program build a 3D model that could be coordinated with the LED location system, then there would be no need for a hand held device. Actually, for large warehouses, any hand held that can connect to the central computer system via wifi/wwan/etc could instruct the system to identify the item. Do you think that it would work outside of the optical range so that people wouldn't be distracted by annoying flashes. I still have to read and attempt to wade through the tech details.
Each item would have unique x,y,z coordinates that would be updated by the scanning system every 1-2 minutes and then each LED on the strip would be assigned a zone of x,y,z coordinates. The LED zone containing the requested item would then activate.
I don't think that building a RF prototype is quite within my budget or expertise but I might be able to manage a computer controlled LED strip. Does anyone have any ideas about making an individually wired, computer-controlled strip? I remember that the good XMAS lights had two wires so that if one of the bulbs blew or detached, the whole string did not die. Does each LED have to be wired individually or is there a way modulate the voltage so that only the nth LED activates? I have no idea if this is relevant but for example, if I want the value of a function to reach 5 at x=3 but nowhere else, I could create the function f(x)= -(x-3)^2+5. I know vaguely that AC follows a trigonometric function but isn't DC different? It would not be too difficult to have an LED only activate when voltage (it is voltage right?) reaches x (in this case 5)
Is it possible to extend the period of the AC wave enough to only have one peak at the desired value?
Thank you for the response
|Sep27-07, 01:25 PM||#8|
There are light-to-pick systems that illuiminate the tray conatining the part to pick, there are even laser scanner based ones which steer a visible laser do to the next bin to avoid having to wire up exisitng bins.
For the LEDs you could look at an I2C bus, this is a simple 1 or 2 wire bus with very cheap individual addressed nodes on it. You could attach an LED to each node and send a command to turn on that LED - in quantity it is likely to be much cheaper than an analog system.
|Similar Threads for: An Idea|
|No Idea What To Do||Introductory Physics Homework||2|