Need help with sensors

tricky, and hard to tune.

you'd have trouble with having the device figure from a fall vs sitting down quickly.

http://www.toptenreviews.com/health/senior-care/best-fall-detection-sensors/

try these

 

Humpty Dumpty?

 

not really, because the acceleration when free falling is at maximum ~ 9.8 (depends on your velocity) so it is pretty easy to know if someone is falling or not.
and you can also set a timer to know at which speed the fall can prove dangerous or fatal
also the acceleration when hitting the ground is the best factor to tell wether or not the fall can prove fatal

 

A gyroscope for change in direction and an accelerometer for the sudden acceleration. They both should do.

And it would be better to ask more specialised sites rather than here. You can try Reddit or Quora. and maybe Google for few more.

 

not really especially if the sensors are on the wrist or hands there are lots of motions that can accelerate and or stop faster than an unplanned fall and cause a false alarm. Especially since people don't fall at 9.8 meters per second (20ish miles per hour)
yes that's the acceleration due to gravity, but you aren't going to accelerate to that speed in the 3 feet or so it takes to "fall"

this is one of those "easy on paper" "hard in reality" problems.

 

I thought you need help censore "stuff" turns out you need help to sensor stuff

 

i don't think the gyroscope is necessary as for as long as you are submitted to a great enough acceleration, it could prove fatal.
for exemple, let's say you've been hit by a car, you won't go down, so what? will the detector not say a thing?
so i think that as long as the acceleration submitted to the sensor is great enough, it should pick it up, no matter the direction. so the gyroscope is unnecessary expenses

 

Life?

 

you know that at low velocity the air resistance is mostly negligible right??

then what? what would cause the body to fall at less then the pull of the earth??

and lile i said, it should be wore near the center of gravity since it won't move much due to your movement

 

3 way accelerometer attached to a belt should do the trick, I once bought one from eBay for real cheap. Next would be to drop test and check the kind of signal generated during a fall. I am assuming it would be a sudden drop to 0 after gradual rise. You might also want to check other situations that could give false positive.
Just make sure to avoid directional sensors like ir and ultrasonic. Too prone to interference in this case.
you can probably try a piezoelectric sensor to detect the shock from falling.

 

If you need it for practical reasons, such as a grandparent, ask your doctor or if they don't know then a medical forum. If you want to develop a product, then you're going to have to figure out how to discern falling from sleeping and the like

 

LOL you think 9,8m/s it the maximum speed to which you accelerate?

 

i'm really curious how you'd detect a fall with a piezoelectric sensor with a 100% accuracy...
also more than the acceleration when falling, the one when you are submitted to a shock is the deciding factorto determine injuries

what was the max velocity of a human body again, something like 50 m/s

 

Hell if I know I'm a law student XD

 

that was a rhetorical question

 

No it's the maximum rate [at] which you accelerate in a vacuum due to the earths gravitational field.

However assuming the average height of a human to be 72 inches (it's not) with the standard .56 height to center of mass ratio for humans. that gives us about .99 meters of distance. Rounding up to 1.

Ignoring air resistance due to the low speed and short distance

assuming the person were to "pratfall" with no leverage or arcs due to the rest of the body interacting with the ground. The individual will hit at a speed of roughly 10 miles per hour, in about 1/2 a second after the fall commences.

Now given that the current world record for a "punch" is roughly 45 miles per hour, which means that is is quite possible for the human body to move faster not falling, than falling. The point is, that for any device you use, especially on the extremities, it's highly likely that a human gesture can trigger a false alarm. Moreso if you want to always detect falls. Since if you decrease the sensitivity to minimize false alarms, you increase the amount of times the device would actually miss a real fall.

 

maybe you should read what you previously said...

what the heck are you trying to say ?

 

sweet tapdancing.....

A.) the human body (in whole or in part) can accelerate and decelerate at rates faster than 9.8M/s^2, which means just relying on acceleration rates alone will not adequately serve to accurately identify falls or not.
B.) on [average] the human body's center of mass from a standing position to a prone position, will travel 1 meter which puts a very rough maximum impact velocity of 10mph. Where the damage is going to drastically depend on the surface the individual falls on, and the conditions of the fall itself. (consider a fall on concrete vs sand for example)

the point is that just an accelerometer alone is not going to make a perfect "fall sensor". There is a reason that even the professional devices when tested miss about 1 out of 5 simulated falls, and still generate false alarms on a fairly regular basis.

All in all the original poster is far better off purchasing a simple, engineered and researched device for about 100 or so. (like this one https://safeguardian.com/products/sos-carecaller-fall-alert-mobile-help-alert) Rather than juryrigging an expensive and minimally functional radio-shack special together.

 

for the A) well that's obvious.. but when you hit the ground, the acceleration felt is especially high, thus it is not difficult to know if there was a chock felt by the individual. but in the case where said individual has fallen through the stairs, the problem can be a bit more tricky and it is true that it won't necessarily trigger an alert due to the fact that the human body is rather soft.
but if you add a gyroscope, it could be possible to know the position of said individual and thus reduce the acceleration required, it's true.

btw only the best of the best can come out unscathed when falling at 10 mph, but do you expect your regular elderly to survive such a fall ?
(roughly 10 meter or 33 feet)
and this :"on [average] the human body's center of mass from a standing position to a prone position, will travel 1 meter" i don't understand why you say this, especially the "will travel", what is the context ? and how do you come to this conclusion :"which puts a very rough maximum impact velocity of 10mph" just maybe, i don't know, explain that the force felt by the individual is the maximum supported before breaking a bone (which is by the way highly dependent on how you fall), or before being unconscious, something like that...

and i'd like to know where you get your statistic of 1 out 5 times...

but i do agree that he should buy something built by professional instead of making its own as in fact, it's less costly, has more function and is foolproof contrary to a homemade one. (also such an item is bound to have false alarm, but well better safe than sorry)

 

Right there in that link, the average fall detector tested out at a 75% detection rate. Even the industry claims a 95% rate (which means a 1 in 20 fail rate)

I'm not sure how i feel about trusting a D20 roll to fall safety.

Basic physics

The human center of mass located at about .56 of the human's height. If the average human is 6 ft tall (they aren't) that put's their center of mass at roughly 1 meter off the ground.

Which means they have at most a vertical drop of 1 meter for their center of mass which gives you a maximum final speed of 10 miles per hour (assuming that you are dropping someone like a brick) [first assume a spherical rigid human]

This is modified of course, by several factors such as the fall surface and if any limbs get involved to brake or slow the fall. As well as the order of impact of the subject. For example if the main point of impact is your chest or limb, you'll probably come out clean as most of the kinetic energy will either be dispersed through the mass or absorbed via the flexibility of the limbs. Of course if you're falling and you wind up hitting a corner to concentrate that impact site somewhere vital, like say, your head. that's a whole nother issue.



For the record, if you check the CDC mortality rates for falling from heights (ladders in this case) located here Your odds of death from a 5 foot fall (higher than what we are talking about with a trip/slip are relatively small. Lower in fact than the odds the sensor wouldn't even detect a fall.

More specifically if we are looking at reducing the risk factors in older people, fall sensors are among the least efficient ways to go about it.

https://www.cdc.gov/homeandrecreationalsafety/falls/compendium.html

As even something as simple as Tai Chi 3x a week dramatically reduced both the fall rates and the injury rates from falls.