|
Firearms Accessories: Holsters, Safes, Lights & more If it locks up, carries, fits on to or cleans up your firearms, discuss it here. |
|
Thread Tools | Display Modes |
#1
|
||||
|
||||
Project: A rail sensor & app to coach your shooting
I have an idea for a project which I'm hoping to start seriously on soon. I mainly want to describe it in case somebody knows why it clearly couldn't work, although other feedback is welcome. I'm not actively seeking assistance building the thing (I think I can do it, sort of), although feel free
The project is a combination of a sensor accessory which would be attached to a pistol rail, together with a simple smartphone app. Basically, you'd go to the range, attach the sensor to your pistol, start the app and prop up your phone so you could see it while shooting, and shoot. Your phone would display messages like "You're anticipating recoil!" or "You're milking the gun!" or "Good shot! That one should be on target!" or something like that. The sensor would basically be a little 3d-printed box that fits on a rail and contains an arduino microcontroller, an accelerometer/gyroscope/compass (AGC) chip, and a bluetooth chip. All it would do is pair with the phone and send the AGC data to the app. The app would also be fairly simple. After spending a lot of time gathering training data, it should be pretty easy to for a program look at the stream of data coming in from the AGC and tell whether a shot was just fired. When that happens, you take the chunk of data from just before the shot (like maybe 0.1 seconds' worth) and see whether it's pretty constant (i.e. the gun hasn't moved) or if it matches one of the patterns of a shooting error. Again, training data would be required to decide how big of a change on one of the sensors (AGC chips typically have 3 different axes for each of the A, G, and C parts respectively) indicates an actual error as opposed to just noise. Then you draw a message on the screen telling the user whether they screwed something up or not. To clarify, the product would do nothing like attempting to determine visually if your shots are actually on target. It just tries to provide more immediate feedback on whether or not you made an error in manipulating the firearm while shooting. Would this thing work? I don't really know if you can get sufficiently good data from hobby-grade AGC chips to classify a time series into errors. Like, the difference between anticipating recoil and noise (on a good shot) might be too small to detect. Or there might be some other issue I didn't think of. I basically have to just build it and do a bunch of shooting with it and then crunch the numbers and see. Still, if it's obvious to anyone here why it wouldn't work, I'd be very interested to hear. Assuming it does work, and well, and people wanted one, I wouldn't really be looking to make a commercial product out of it. Basically, I'd give away all the code and data and design stuff, and let random companies build and sell compatible sensors. Or you could just roll your own; the chips you'd need are about $50, although I'd guess some kind of ASIC board would be a lot cheaper than the prototyping chips I'd be using. Thoughts? Criticism? Personal abuse? |
#2
|
||||
|
||||
|
#3
|
||||
|
||||
****, never heard of that. Yes, that's exactly what I was going to make. Guess I didn't try hard enough to find an existing implementation when I first started thinking about it. Which was kind of stupid because my main thought was "This is a good idea, why hasn't anyone done it yet?"
I suppose this at least answers the question of whether it's doable. I doubt that $150 price tag is to cover the cost of high-grade AGC chip (rather than just provide extra profit). Maybe I'll make my own anyway and post details. Anyway, thanks. |
#5
|
||||
|
||||
I hadn't heard of it, but the principle makes sense. It was just a question of whether you could get a sufficiently good signal.
The thing that impresses me about the mantis is that they claim that you can use it to practice dry-firing. That strikes me as hard to do, especially without lots of firearm-specific tuning. But they do note that when doing dry-fire practice, you may get false positives. |
#6
|
||||
|
||||
they're a sponsor of active self protection on YouTube. I never would have known about it otherwise. like you said the concept seems sound. I may pick one up somewhere down the line to maintain skills.
|
#7
|
||||
|
||||
Great idea. But don't be discouraged that someone thought of the general concept already. It just validates that it was a good idea. If you keep at it long enough, you'll hit a home run.
Cool post. Welcome to calguns. Too bad I'm not still in the bay area. |
#8
|
||||
|
||||
tenemae: thanks for the encouragement. It definitely seems like a less urgent thing to do since somebody's already done it, but I may well do it anyway.
One thing that I think would be cool to do would be to make it an app for Google Glass, or something else whereby you could get more immediate feedback than on your phone. People basically say that the problem with the MantisX is that you basically shoot your mag, then get feedback from your phone. If you were wearing glass, you could have feedback after each shot. This would be something my device (my working name was ShooterTutor) would do that MantisX didn't, although it'd be easy enough for them to add. Another option might be a smartwatch. You could wear your smartwatch on the inside of your wrist and be able to see very simple feedback info (such as color codes or big arrows) after every shot. Then after you empty your mag, you could look at your phone for more precise info. |
Thread Tools | |
Display Modes | |
|
|