Announcement

Collapse
No announcement yet.

Punch Sensor Prototype

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Dr_Awesome
    replied
    Originally posted by ChenPengFi View Post
    Regarding "Pain receptors"...

    There is no such thing.
    "Nociceptors" are receptive to heat, tissue distortion and differences in pH.
    "Pain" isn't a measurable "thing" and is an output of the brain.

    http://www.bullshido.net/forums/showthread.php?t=121734
    Shit.

    That was one of my favorite threads on here, by the way.

    I need to stop using that language, and I noticed I used it in a bullet point on the video too. It feels more efficient when I say it, but it paints a completely wrong mental picture of there being some special switch under the skin that makes you feel pain instead of the bigger picture where information is sent and output is based on context.

    Thanks.

    Leave a comment:


  • Dr_Awesome
    replied
    Originally posted by W. Rabbit View Post
    The more I go over physical sensor-based DSP systems, I the more I think you should digitize your signal before feeding it to the MC. It'll solve your Arduino protection problem, and provide a cleaner signal into it that can be more easily sampled and you'll lose very little analog signal, if any.

    ADCs are commonly used as pre filters for these kinds of sensor systems when the desired end result is a signal analysis.
    http://playground.arduino.cc/CourseWare/AnalogInput

    Leave a comment:


  • Dr_Awesome
    replied
    Originally posted by cereus View Post
    Now that it's been spelled out to me I am very interested in seeing where this goes. I never felt like force ever gave a clear picture.

    Can you give examples of what punches (perhaps to which targets) might require more "push" or "snap". I haven't done any striking for the past 10 years and tkd from 8-14 didn't really give me any real education in the matter.

    Leave a comment:


  • Dr_Awesome
    replied
    Originally posted by hungryjoe View Post
    Dr_Awesome,

    Are you using an accelerometer chip in your setup?
    No. For now, I am using the speed of compression of the crystals to measure the velocity, but I have been toying with the idea of using both, because right now I cannot tell if that is part of the reason for my high error rates, or if it is just the build itself. I picked up a cheap accelerometer to test out just in case, so maybe I will use it in next generations...

    Leave a comment:


  • ChenPengFi
    replied
    Regarding "Pain receptors"...

    There is no such thing.
    "Nociceptors" are receptive to heat, tissue distortion and differences in pH.
    "Pain" isn't a measurable "thing" and is an output of the brain.

    http://www.bullshido.net/forums/showthread.php?t=121734

    Leave a comment:


  • W. Rabbit
    replied
    The more I go over physical sensor-based DSP systems, I the more I think you should digitize your signal before feeding it to the MC. It'll solve your Arduino protection problem, and provide a cleaner signal into it that can be more easily sampled and you'll lose very little analog signal, if any.

    ADCs are commonly used as pre filters for these kinds of sensor systems when the desired end result is a signal analysis.

    Leave a comment:


  • cereus
    replied
    Now that it's been spelled out to me I am very interested in seeing where this goes. I never felt like force ever gave a clear picture.

    Can you give examples of what punches (perhaps to which targets) might require more "push" or "snap". I haven't done any striking for the past 10 years and tkd from 8-14 didn't really give me any real education in the matter.

    Leave a comment:


  • hungryjoe
    replied
    Dr_Awesome,

    Are you using an accelerometer chip in your setup?

    Leave a comment:


  • W. Rabbit
    replied
    You can try digitizing using an ADC, and feed the Arduino digital in values from that in instead of analog current/voltage directly from the sensor.

    http://www.maximintegrated.com/app-n...ex.mvp/id/3775
    Last edited by W. Rabbit; 12/30/2013 9:22pm, .

    Leave a comment:


  • Dr_Awesome
    replied
    Originally posted by cereus View Post
    So tell me, as an MA noob and borderline Luddite (through laziness if I'm honest), what exactly would you like your punch sensor to tell us that other sensors don't?
    Momentum Transfer and energy transfer.

    Momentum transfer tells you how much you push your opponent (or parts of your opponent, like his head or his hands or the air in his lungs).

    Energy transfer is associated with structural damage to the local tissue and pain.

    Most sensors measure "force" because it is easy to measure, and in colloquial english the word "force" sounds like an awesome things to measure for a punch, but the actual physical quantity is not helpful to us unless we plot the whole curve over time and then translate that curve into momentum and energy.

    I tried to spell out some of these distinctions in a prior blog entry here:
    http://theamazingdoctorawesome.blogs...our-fists.html

    Leave a comment:


  • cereus
    replied
    So tell me, as an MA noob and borderline Luddite (through laziness if I'm honest), what exactly would you like your punch sensor to tell us that other sensors don't?

    Leave a comment:


  • hungryjoe
    replied
    Originally posted by Dr_Awesome View Post
    I've seen a few other different, either in the form of gloves or targets, and they almost always measure "force" which makes no sense to me. "Force" is easy to measure (your bathroom scale is an excellent force meter), but a punch has a very rich force curve over time, and whether they measure it by "peak force" or "total force" it still doesn't tell you anything important about a punch. Someone could easily lean on you and apply more "force" than you would feel with a knockout punch. Do you remember what those sensors did if you leaned on them or just pushed them with your hands? Did it end up teaching you to just push really hard after impact each time?

    Directionality is a tough issue to deal with, which was why I tried using three sensors in parallel, but it is tough to make them identical when you do it by hand.

    I really cannot stress how much you need to measure both energy and momentum for it to make any sense...
    Point being, F=ma equals nothing if you're not on target.

    Enjoying the read. Looks as though you're moving forward.

    Leave a comment:


  • Dr_Awesome
    replied
    Originally posted by hungryjoe View Post
    Century Martial Arts had a sensor integrated into a bag around twenty years ago. I don't remember seeing it in a catalog. My instructor at the time worked there, so perhaps it was a prototype.

    Being off target just a bit changed the reading dramatically.
    I've seen a few other different, either in the form of gloves or targets, and they almost always measure "force" which makes no sense to me. "Force" is easy to measure (your bathroom scale is an excellent force meter), but a punch has a very rich force curve over time, and whether they measure it by "peak force" or "total force" it still doesn't tell you anything important about a punch. Someone could easily lean on you and apply more "force" than you would feel with a knockout punch. Do you remember what those sensors did if you leaned on them or just pushed them with your hands? Did it end up teaching you to just push really hard after impact each time?

    Directionality is a tough issue to deal with, which was why I tried using three sensors in parallel, but it is tough to make them identical when you do it by hand.

    I really cannot stress how much you need to measure both energy and momentum for it to make any sense...

    Leave a comment:


  • Dr_Awesome
    replied
    Originally posted by W. Rabbit View Post
    I had a couple of thoughts.
    What is the Ardruino's sample rate for analog in? If it's not high enough and without other components you could be losing information. I doubt it's being lost on the dummy side of the circuit. :)

    Leave a comment:


  • Dr_Awesome
    replied
    Originally posted by W. Rabbit View Post
    I had a couple of thoughts.

    First, LEDs!

    Second, I noticed all your graphs were in the time domain, but you are generating AC in the frequency domain (crystals!).

    What is the Ardruino's sample rate for analog in? If it's not high enough and without other components you could be losing information. I doubt it's being lost on the dummy side of the circuit. :)

    Leave a comment:

Collapse

Edit this module to specify a template to display.

Working...
X