Targeting Computer

azredhawk44 · March 18, 2010, 04:42 PM

I've been reading up heavily on Arduino boards, which are little hobbyist microcontrollers capable of being programmed by an end user to take input from various sensors and then use user-programmed logic to either generate output, or make a mechanical action (via a motor or other device attached to a circuit on it).

People make self-deterministic yet simplistic robots out of them.

Here's some, playing "soccer."

http://www.youtube.com/user/TeamVossius

I'd like to try and build a self-deterministic robot that can out-shoot me. Elements would include:

- potential target detection via a NTSC camera mounted at the focal point of a scope
- target analysis to make certain the target conforms to expected standards
- range finding capability
- temperature and barometric sensor analysis
- wind speed and direction analysis
- ballistic computation
- mechanical adjustment to compensate for above factors
- a firing circuit interrupt that mechanically breaks the circuit between the board and the motor/solenoid for the trigger actuation unless the user assesses the safety of the shot and closes the firing circuit interrupt
- ability to fire a semiautomatic rifle, wait for recoil and then assess environmental factors again for a new shot, then fire once the decision makes logical sense and a new "all clear" is delivered via the firing circuit interrupt.
- eventually, I'd like to have "it" entered into a High Power Match or similar environment, and I would program it with appropriate authorization for the number of shots in a given string.

A couple concerns:

#1 (The Big One): Would this be a machine gun? Would the definition of a "trigger operation" be simply turning the computer on? Can a person get a ruling on this from the ATF tech branch without actually building it and sending it to them for a ruling? Can such a project be acceptably approved for research purposes, either possibly by DARPA or by the ATF tech branch?

#2: No reason for a shoulder stock on such a device. Would such a device need to be licensed as an SBR? I'm thinking the final device would end up looking like the .50 caliber platform from the movie "Shooter."

Obviously the project is very complicated and would be accomplished in a series of small steps, starting probably with a simple tripod robot with a laser pointer to "fire" rather than an actual gun, probably working on incredibly simplistic black/white targets at ranges from a foot to perhaps 10 feet. That would probably keep me busy for at least several months. Long range target detection may end up being beyond the processing capabilities of an Arduino board. I'd be playing with a simple laser pointer for quite some time before I got ready to strap an actual rifle to such a platform.

Terminator/Matrix logic would be left out, I think.

Arduinos are only 16mhz computers. Pretty small.

Jim Watson · March 18, 2010, 04:56 PM

Put the first actual firing model on a single shot rifle to avoid any hint that you are building a MG. I think a computer directed followup shot capability would scare the feds into defining it as a full auto. That will leave you plenty of work to do on the Bob Lee Swaggert I-Sniper.

You could not actually enter it into a match, self aiming scopes and rifles are specifically banned in the rules. Maybe you could set up a head to head competition like the Deep Blue chess computer games.

mapsjanhere · March 18, 2010, 06:16 PM

Usually the definition is something like "one action, one shot". So if you press the button five times, and the computer decides to shoot 1 - 5 shots, that MIGHT get past the FA rule. But in addition to the NFA I would also check my wildlife/hunting rules, such a thing might easily be mistaken as a poaching device, the high tech version of a shotgun with buckshots and a string across a deer trail.
You would probably have less to worry about the SBR as long as your barrel and overall length meet the requirements.

Suwannee Tim · March 18, 2010, 06:44 PM

I looked at the specs for the micro-controller, it has no DSP capabilities and the CISC architecture is ill suited for DSP applications. The low bandwidth of the ports and the fact that they are controlled directly by the micro-controller, not by a UART or similar device makes this device particularly ill suited for image capture and processing. I think you will need a more powerful computer. A vastly more powerful computer in fact, one that has hardware purpose designed for real-time image acquisition and processing. As far as the BATF, why don't you send them a letter asking these questions? The little bit I know about Class III, I know that you cannot get a permit to do "research" that enables you to build a machine gun. Otherwise, I would be doing "research".

BitterTait · March 18, 2010, 06:59 PM

Someone's done something similar, although simpler and with a pellet gun:

http://www.whazsup.com/Tinker/blog/?p=401

There's also an annual automated airsoft contest at

You'd need more iron than an arduino can provide for visual detection; it's not the speed, it's the memory and space requirements.

As far as your questions go, I wouldn't worry about #2, as long as the whole thing is longer than the minimum length. Look into some of the so-called "rail guns" that are used in the long range unlimited category, there's no stock, it's just an action and a barrel mounted to a chunk of iron.

I'd skip dealing with question #1 by having it asked by having it only fire when someone presses a button. This has the advantage that someone confirms each shot before it fires, reducing the chance of mishaps. If you take your device to the range, you won't have a problem finding someone to run it at the range, bring a baby like that out and half you're range time will be taken by showing it off.

LaserSpot · March 18, 2010, 07:15 PM

Quote:

the high tech version of a shotgun with buckshots and a string across a deer trail

Yeah, if you design it fire autonomously, I think it would be classified as a destructive device, like a land mine or a spring-gun: http://en.wikipedia.org/wiki/Spring_gun

It would be better to make it work like a point-and-shoot camera; you wait for the light to turn green, then press a button to shoot.

We're already fighting a war in Pakistan using armed, semi-autonomous drones. This technology is the next "Revolution in Military Affairs", according to a book I'm reading:
http://www.amazon.com/Wired-War-Robo.../dp/1594201986

Dr. Strangelove · March 18, 2010, 07:36 PM

I would contact the BATF, tell them what you would like to do, and let them provide recommendations on how to proceed.

You may well find there is more "official" interest in a "good" way in something like this than you may first think...

March 18, 2010, 06:16 PM	#3
mapsjanhere Senior Member Join Date: August 6, 2009 Location: Albuquerque Posts: 2,832	Usually the definition is something like "one action, one shot". So if you press the button five times, and the computer decides to shoot 1 - 5 shots, that MIGHT get past the FA rule. But in addition to the NFA I would also check my wildlife/hunting rules, such a thing might easily be mistaken as a poaching device, the high tech version of a shotgun with buckshots and a string across a deer trail. You would probably have less to worry about the SBR as long as your barrel and overall length meet the requirements. __________________ I used to love being able to hit hard at 1000 yards. As I get older I find hitting a mini ram at 200 yards with the 22 oddly more satisfying.

March 18, 2010, 06:44 PM	#4
Suwannee Tim Senior Member Join Date: June 27, 2009 Location: Jacksonville, FL Posts: 320	I looked at the specs for the micro-controller, it has no DSP capabilities and the CISC architecture is ill suited for DSP applications. The low bandwidth of the ports and the fact that they are controlled directly by the micro-controller, not by a UART or similar device makes this device particularly ill suited for image capture and processing. I think you will need a more powerful computer. A vastly more powerful computer in fact, one that has hardware purpose designed for real-time image acquisition and processing. As far as the BATF, why don't you send them a letter asking these questions? The little bit I know about Class III, I know that you cannot get a permit to do "research" that enables you to build a machine gun. Otherwise, I would be doing "research". Last edited by Suwannee Tim; March 18, 2010 at 06:58 PM.

March 18, 2010, 06:59 PM	#5
BitterTait Senior Member Join Date: December 9, 2009 Location: Michigan Posts: 144	Someone's done something similar, although simpler and with a pellet gun: http://www.whazsup.com/Tinker/blog/?p=401 There's also an annual automated airsoft contest at You'd need more iron than an arduino can provide for visual detection; it's not the speed, it's the memory and space requirements. As far as your questions go, I wouldn't worry about #2, as long as the whole thing is longer than the minimum length. Look into some of the so-called "rail guns" that are used in the long range unlimited category, there's no stock, it's just an action and a barrel mounted to a chunk of iron. I'd skip dealing with question #1 by having it asked by having it only fire when someone presses a button. This has the advantage that someone confirms each shot before it fires, reducing the chance of mishaps. If you take your device to the range, you won't have a problem finding someone to run it at the range, bring a baby like that out and half you're range time will be taken by showing it off. __________________ Powered by VI!

March 18, 2010, 04:42 PM	#1
azredhawk44 Junior member Join Date: September 28, 2005 Location: Mesa, AZ Posts: 6,465	Targeting Computer I've been reading up heavily on Arduino boards, which are little hobbyist microcontrollers capable of being programmed by an end user to take input from various sensors and then use user-programmed logic to either generate output, or make a mechanical action (via a motor or other device attached to a circuit on it). People make self-deterministic yet simplistic robots out of them. Here's some, playing "soccer." http://www.youtube.com/user/TeamVossius I'd like to try and build a self-deterministic robot that can out-shoot me. Elements would include: - potential target detection via a NTSC camera mounted at the focal point of a scope - target analysis to make certain the target conforms to expected standards - range finding capability - temperature and barometric sensor analysis - wind speed and direction analysis - ballistic computation - mechanical adjustment to compensate for above factors - a firing circuit interrupt that mechanically breaks the circuit between the board and the motor/solenoid for the trigger actuation unless the user assesses the safety of the shot and closes the firing circuit interrupt - ability to fire a semiautomatic rifle, wait for recoil and then assess environmental factors again for a new shot, then fire once the decision makes logical sense and a new "all clear" is delivered via the firing circuit interrupt. - eventually, I'd like to have "it" entered into a High Power Match or similar environment, and I would program it with appropriate authorization for the number of shots in a given string. A couple concerns: #1 (The Big One): Would this be a machine gun? Would the definition of a "trigger operation" be simply turning the computer on? Can a person get a ruling on this from the ATF tech branch without actually building it and sending it to them for a ruling? Can such a project be acceptably approved for research purposes, either possibly by DARPA or by the ATF tech branch? #2: No reason for a shoulder stock on such a device. Would such a device need to be licensed as an SBR? I'm thinking the final device would end up looking like the .50 caliber platform from the movie "Shooter." Obviously the project is very complicated and would be accomplished in a series of small steps, starting probably with a simple tripod robot with a laser pointer to "fire" rather than an actual gun, probably working on incredibly simplistic black/white targets at ranges from a foot to perhaps 10 feet. That would probably keep me busy for at least several months. Long range target detection may end up being beyond the processing capabilities of an Arduino board. I'd be playing with a simple laser pointer for quite some time before I got ready to strap an actual rifle to such a platform. Terminator/Matrix logic would be left out, I think. Arduinos are only 16mhz computers. Pretty small.

March 18, 2010, 04:56 PM	#2
Jim Watson Senior Member Join Date: October 25, 2001 Location: Alabama Posts: 18,535	Put the first actual firing model on a single shot rifle to avoid any hint that you are building a MG. I think a computer directed followup shot capability would scare the feds into defining it as a full auto. That will leave you plenty of work to do on the Bob Lee Swaggert I-Sniper. You could not actually enter it into a match, self aiming scopes and rifles are specifically banned in the rules. Maybe you could set up a head to head competition like the Deep Blue chess computer games.

March 18, 2010, 07:36 PM	#7
Dr. Strangelove Senior Member Join Date: August 1, 2008 Location: Athens, GA Posts: 1,436	I would contact the BATF, tell them what you would like to do, and let them provide recommendations on how to proceed. You may well find there is more "official" interest in a "good" way in something like this than you may first think...