I’m basically trying to make an electronic attendance machine to count the number of students in a class and so that they can give their respective roll no’s and not give any one elses by mistake or intentionally.
I plan to go about making it based on the Intel 8051 chip. The problems I face are one or more students might try to give attendances for friends which is to be prevented. The count is to be calculated when the student enters the class room . The professor then enters his code for his specific tutorial. The students then enter thier unique code ( by way to keying in or by barcodes) and they registers thier attendance.
Problems
1 - How do I catch the culprits
2 - How do ensure that there has been proper count of the students entering the class ( I thought of using lasers but it doesnt seem to fit in ie ppl can walk past it a couple of times registering a false count or if someone leaves the class the count still gets registered (for which I figured out using two lasers and the count gets registered according to the laser that is deactiveted or activated first , but that is also not foolproof if you consider a person who is wildly swinging his arms and makes the device beleive someone is entering the class)
Any and all ideas are welcome.
Thanx
Jay
PS : If you have any questions regarding my queries let me know!
What ever happened to good old fashioned being there in person to watch them?
I suppose you could give the students RSA SecurID tokens and have them log into something. Some of the core tenets to SecurID are the user has to phyically have the token and mentally know the PIN. (Not to mention how to put them together.)
Ehhh… scratch that. The tokens are a little expensve and somewhat fragile. Besides that, what’s to stop the miscreants from giving their PIN and token to someone else to enter?
Unless you use biometrics, there’s not much you can do on an automated basis to ensure that someone is who they say they are.
Magnetic key cards that read the students ID as they enter the room. Set sensitivity level so that card key will read anywhere on body, not requiring it to come out of bag or pocket.
Instruct students to enter room one at a time.
Set alarm so that if two card keys are sensed at the same time, a laser comes on and shoots the offender in the eyes. To find the offending student, just look for the one running into open lockers.
Perhaps you should require the students to have barcodes tatooed onto their foreheads, so a scanner can take their attendance when they come in.
On a more serious note, I think you’re better off with face recognition software. Don’t get me wrong, for picking out a handful of known terrorists from amongst the thousands of people who travel through an airport, face recognition is horrible. But to tell which of your 100 students is entering a classroom? It’s up to that. (“Which of these 100 people is this person?” is a much easier problem to solve than “Is this person one of these 10000 people?” Of course, it’ll scew your results if you’re worried about other people (from outside your class) taking in the class instead of the people who are supposed to be in it)
On a different note, this whole idea is fairly orwellian. If you can’t tell at exam time which students attended your lectures, then you either give bad exams, bad lectures, or the student was bright enough to learn the material on his own and shouldn’t be punished for not wanted to sit through your lecture.
A cheapo webcam taking a picture each time a code is punched in would be interesting. Wouldn’t have to actually be functioning, just look like it was. Light comes on, goes off etc.
Avoid lasers. Your friendly school lawyer will explain why.
I have seen fingerprint recognition systems for quite cheap. Any other biometric wouldn’t be cost effective in your situation.
The fingerprint system sounds feasable. How about the low-tech method of having them sign in? Each day they sign their names in the same numbered space, check for discrepancies. I doubt people would be able to forge another’s signature effectively enough for a side by side examination. If so, you have some very sneaky students, who will most likely defeat other, more expensive, measures.
Is the system you’re designing just for one class, or for a whole campus? Is there an existing ID or access control infrastructure you can tap into? For example, do the students already have a card or badge with machine-readable data on it?
What degree of administrative overhead is acceptable? For example, even if your system is based on a unique code per student, you’ll need some way of assigning the codes, dealing with people who forget their code, etc… The issues are similar if you need to manage cards or other physical tokens or register biometric signatures.
How are you going to deal with exception cases? Let’s say I show up to class, but am too hung over to remember my code…or I lost my wallet with my card…or I got in a fight last night and my face is too swollen for my facial geometry to match…or perhaps I had a freak mandoline accident and grated my fingerprint off. Do I still get credit for showing up?
What environment is this intended for? High School? University? Testing centers?
Is it intended for daily use, or only for exams?
Are there regulatory requirements the system needs to satisfy or comply with?
How large is the average class? The largest class?
In general, does the instructor know the students well enough by sight to be able to identify them?
Are you doing this in real life, or is it just an academic exercise?
Why are you thinking of making a special-purpose machine rather than using a general purpose computer?
It’s basically a project I have to make for my engginering graduation. As of now it’s bascially for just one class but that should not be a limitation. For the coding system we are going to integrate with a microcontroller ( Intel’s 8051) and hopefully we can assign codes to everyone in it’s software. As for the exceptions cases : Haven’t thought about that but eventually I’ll have totake that into account , if my final design includes a code for everyone that shoundnt be a problem ( If they forget there goes the attendance)
It is intended for daily use . There are non such requirments it has to satisfy . If you could be a little more specific on this part I could clarify it. As I said it’s for my project and should be able to use it in real classroom life.
Use an anti theft device from a department store, you know, those things by the door that you have to walk thru to get out the door. the student could carry the little metal tag anywhere on their body, wallet, purse, pocket, however. Maybe the tag would have unique info on it for each student, By arranging two detectors in a row you could tell if that student was coming or going.
Ahhh… A final exam of sorts. That would explain your desire to re-invent the wheel.
Well, they’re really, really elaborate wheels, but already done, certainly. (Looking around here at the biometric mantraps and the Motorola Indala badge and RSA SecurID token hanging around my neck.)
With that $100 budget, you’re going to have to have a lot of virtual hardware - SecurID tokens cost us about $55 each, and I have no idea what the Indala RFID badge readers go for. In either case, they can be replaced with keyboard entry - rather than tag an RFID card, the user simply types in their ID number.
If circumventing “social engineering” is high on your list (in other words, preventing someone from entering another person’s ID) you’ll need the biometrics - fingerprint readers are getting pretty cheap.
Ahhh… A final exam of sorts. That would explain your desire to re-invent the wheel.
Well, they’re really, really elaborate wheels, but already done, certainly. (Looking around here at the biometric mantraps and the Motorola Indala badge and RSA SecurID token hanging around my neck.)
With that $100 budget, you’re going to have to have a lot of virtual hardware - SecurID tokens cost us about $55 each, and I have no idea what the Indala RFID badge readers go for. In either case, they can be replaced with keyboard entry - rather than tag an RFID card, the user simply types in their ID number.
If circumventing “social engineering” is high on your list (in other words, preventing someone from entering another person’s ID) you’ll need the biometrics - fingerprint readers are getting pretty cheap.
ftg’s webcam idea is brilliant, IMO. Nothing like a little misdirection (or is it…?) to make people be honest.
Another, similar approach: just have them type in their names (or select them from a drop-down list, etc.) and, when they hit ENTER, have the cam take a snapshot. Then the instructor can review past pics of the students to check for any discrepancies (would just need a program that displayed the last several snapshots of each student side-by-side). Since you’re talking about verifying attendance (and not restricting access–the usu. reason for secure ID tags, etc.), the prof. could review the snapshots at her leisure; even at the end of the semester, if she wished (save every day’s snapshot).
I agree with the others, though–without biometrics (analyzed by either a computer program or a human brain), I think you’re going to have a tough time making this work.
With the 8051, you are going to have to drop any ideas about face recognition.
You need a simple ID unit that the user can carry around. Either a plug in ID (serial interface eeprom on a little PC board with the hardware dunked in epoxy) or a barcode reader.
The eeprom units will be a little more complex, but the reader is much simpler - just a socket to plug into and a little antistatic protection circuitry. The tag unit is REALLY simple. Nothing more than a serial interface eeprom (programmed with an ID number, which can be really long if you are using a 128byte eeprom) and a few traces and some connector pads. If you choose the right connector on the reader, then all you need is some properly sized and placed pads (preferably gold plated, but since you’re on a budget you can just leave the copper and make the students polish the contacts when it doesn’t work right.)
The other way is to find a CHEAP barcode reader with a serial interface. Then you issue a laminated card with a barcode on it to each student in the class.
In either case, your reader collects the codes and passes them onto a PC via a second serial connection - after each class I expect. I don’t expect you want to try to make the 8051 compile records for the whole semester.
As for not cheating, it gets hairy. The only way to really stop cheating is to make the students manually enter a password that is somehow tied to data that they will not want anyone else to have. Something like a credit card number (though I don’t know how common those are among college students in India.) Then again, if you do that you’ve got confidential data for the students that you have to be careful with.
The biometry stuff sounds good, but I’m not sure that the manufacturers are all that wild about giving access to the control protocols to other people. If you can get an inexpensive one with a serial intefrace and you can get access to the protocols, then each student would have his own “ID unit” right there in his fingerprints.
You’ve got a lot of trade offs to make:
An eeprom unit has a simple and cheap reader, but the units are (relatively) expensive.
A barcode unit has a more expensive reader, but the ID units are dirt cheap.
A fingerprint reader would have the most expensive “reader,” but would give you the most certainty that each student was actually there - and you’ve got zero cost for the “ID unit.”
Does it have to count at the beginning of class, or can it count attendance as the students leave?
See, what you do is have the instructor give out a numeric code at the beginning of class. At the end of class as students leave they enter in the given code plus their own unique code. That would solve some of the logistics problems.