midibox Simon says for martial arts

[lief138]

Hi Folks, I am trying to make a midibox for my buddies martial arts teacher.

The purpose of this contraption would be as follows

1) A series of ‘pads’ with internal switches for punching or kicking

2) A console that can program a user defined sequence of pads to hit.

3) A user controlled time that would indicate how long the pads have to be hit

4) A start button on the controller that would start the timer

5) Each pad would have some light to indicate it is the next one to be hit

6) 2 sounds (or perhaps two separate lights) to indicate if the pads were hit on time and/or if the correct pad was hit.

7) Bonus a nice timer display that will have the reaction time for bragging rights

8) Double bonus storing and exporting data for future reference (like in spreadsheet)

9) Triple bonus Pressure sensing pads and some form of readout that indicates force applied

Making the hardware seems easy enough. Core, DIN, DOUT and  maybe an LCD for the console.

Bugfight suggests FSRs for the pads.

The software also doesn’t seem like it would be too complicated.

I think the main difficulty would be to make the pads with an appropriate switch or FSR.

Bugfight also mentions that this device probably already exists in some form. However, This teacher and myself are keen to make this a midibox project.

So anyone got an idea of what this pre-existing device might be so I can research.

So anyone got an idea of how I might make the pads

So anyone got an idea of any thing I might have missed/wanted to add

So anyone got an idea of any problems I might encounter

So anyone got an idea of why am I a Rasta Man

OK thanks

Lief

[stryd_one]

I started writing up a big reply, and it kept expanding with "... if it's X martial art; but if it's Y martial art...."

Which MA are we talking about here?

[nILS]

For the pads you might wanna look into piezos as used in electronic drums (see edrum.info for ideas).

Problems you might encounter: Where to put the indicators so they don't get smashed/hurt the person training ;)

[lief138]

Ok thanks for the feedback.

Somehow it does not surprise me that stryd asks about the specific martial arts style. Its mainly for some semi traditional tae kwon do.  No wing chun wooden man fighting dummies needed. 

I looked at the piezo triggers from edrum. That looks like it could work. I guess some drummer would bash these things pretty hard. I will try to make a trigger and get that working as the first step.

[stryd_one]

Heheh thanks man... I think? :D

How to measure the strikes was what was really making me go in circles. My Tkd experience is limited but I've done quite a bit of Hap Ki Do, which is partially derived from it,  so hopefully I can apply it :/

I know the theory is the same regarding quick recoil in strikes. I imagine your master would probably carry an interest in the amount of time the contact was made for, as well as how hard it was (you don't need to kick someone hard to knock them from a horse, but if you don't take your foot back in a hurry they may decide to use their sword and take your foot with them to the ground.... among other practical reasons)

One of the things that was tricking me a bit, was that using switches to time the strike could easily result in false hits, as could peizos... I'm not sure if your Master teaches a method of blocking where the blocking hand will rest against the body (usually derived from sword practice, you don't want to have the back side of one whack you from a distance) or how much throwing/grappling you're doing, but that could be an issue...

But peizos and FSRs bear some nice properties, like peizos are nice and thin and can be quite sensitive. But then again they don't really tell you if they're being pushed or pulled or just shaken about.  FSRs will only register pressure, so there's much less chance of false hits, but without a very stable surface to mount them on (iie, not a bouncy human) the readings will vary too much to be very useful as far as a "how hard did she hit him?" measurement. this pdf has some good info on those limitations.

I dunno if it sounds kinda crazy, but it seems to me that all three could be a nice combination, or at least two: the peizos and fsrs combined will mean that fsrs will indicate inward force (ie probably a hit) and that can be used to arm the recording of the corresponding peizo for a more useful shock/force measurement. You'd need to have a kind of sandwich assembly for the FSRs to do the trick, but the peizos are well suited to that too. That still leaves the possibility of an accidental trigger by the wearer of the sensors, by blocking, or being pushed, or entering a grapple etc, which is where the switches could be handy, as they can also be used to exclude any 'hits' which are too long, as well as too short....... if you hit someone, you probably lose contact with them within a hundred or so milliseconds, but if you try to drag/push them around, you will usually hold on for at least double that...

Because you're not doing something like dim mak, you won't need quite so many sensors, which is good news; if you can scan the FSR/peizo pairs fast enough, you might be able to use the transitions of the FSR instead of real switches. That'll depend greatly on the particular FSRs you source. MIOS will read 8 AINs every 100us, so I think you could run a lot of sensors in theory... it's just a matter of finding the FSRs with the right response curve, so they're useful as a switch ...

Sorry that's a bit ramble-ey... You can imagine what it was like before I had some direction ;) Hope it helps!?

[lief138]

Hey Stryd, I want to take a step back because I want your feedback but I think we have a misunderstanding.

I wasnt thinking that the user would wear the pads themselves. The plan was to have a target to hit in a given time frame. I spoke briefly with the instructor yesterday and he invisioned at least a 3X3 grid of pads on a surface (perhaps a shhet of plywood) that could be bolted to his wall.

We would want to be able to program the sequence, set the time in to the CS and hit start.

Then the user would have say 1 second to hit the target before it is no longer valid, repeat on next target, etc. Ideally he would like to have a system that generates a score based on the number of valid targets hit and the time taken to hit them.

The idea there being that if you hit the target quicker you would score higher. Also, as soon as one target was hit the next would become active.

So far he does not feel to concerned with measuring the force of the impact although he does say it sounds interesting. from what I have read about the piezos and vibrations, I would probably need to do somethign to isolate each from the vibrations of the board. Perhaps somethign could be done on a software level to prevent false triggers from being counted.

So does this make more or less sense?

[stryd_one]

Sorry man! Obviously I thought this was the type where the targets are worn... LOL, Missed by *that much* :D hahahaah

If your instructor isn't keen on measuring force, I'd be going for some kind of reed switch, probably roll-your-own. Especially if you can find some nice spring steel for the switch actuator, and suspend the pads above the contact point, they should last forever, and you won't have concerns about isolation of peizo triggers - which is gonna be darn-near impossible if you're tkd-kicking a stationary frame, and would probably be detrimental given that you want to know if the person half-misses the target and catches the neighbouring pad in addition to the intended one (or if they just plain miss it entirely). If measuring force does become a goal later on, then I'd be going for FSRs, given that you now have a nice stable platform on which to mount them.

Although you can get some very thick and strong impact grade plywood, I probably wouldn't use something like that as a backboard for this, because it tends to splinter quite badly when it does break, and you wouldn't want to pass a foot/leg through a hole like that. I'd probably stick with solid timber for the frame, and use the ply only for the pads themselves. That being said, I'd design the whole thing so that it can be replaced in parts... TKD players, play rough ;)

One of the ideas in my big draft was that you could use an LTC module to send data in realtime from the pads to a PC/laptop/etc, which could give some nice features: As well as logging data (which could also be done direct to a bankstick) and visualising it nicely (which could also be done to a lesser extent with a large 7segment display), you could also do some programming in realtime, (inserting delays, turning on 'moving targets', other tricky buttons, or even just queueing up predetermined sequences) which could be handy because you can expect the yudansha to try and cheat the system ;) ... at least now, I don't have the concern that the person wearing the targets would have to wear a 'tail' like in fencing competition: those things are a PITA.

Am I heading in the right direction this time? :D

[nILS]

We once made something relatively similar with piezos and the small pads from a used "hantelbank" that we got off a blowout sale of a local sports supplier. sth like this:

t

Basically we made diy pads like the edrum people with a cushion in front of them ;-) I just asked the guy and it still works. He's using it as play controls for his studio (after he broke a whole bunch of controllers he wanted something sturdy).