Fully Autonomous Sentry w/Target Anticipation

[David08]

When is defcon going to have a comp for sentries that track and anticipate moving targets. the whole 'shoot at 10 stationary targets of color x in super controlled environment' idea is pretty basic. Here's a video of one i built back in April that i guarantee (if you haven't already seen it) performs better than anything you've seen. I've made many improvements since then (they can be seen on the website[listed in the video description]) but even that version is better than anything else out there that i know of. i posted this hoping there was something out there that could compete with this (specifically in performance, portability, and targeting) and to hopefully get some more intelligent feedback than the typical "it could shoot the cats in my yard" or "spy sappin mah sentry!" crap I've heard over and over.

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

Here's a video of one i built back in April that i guarantee (if you haven't already seen it) performs better than anything you've seen.

I think this one performs better: http://gizmodo.com/gadgets/tag/robot...you-212241.php

Also there is a thing that is called the MK 15 Phalanx CIWS. I'm pretty sure that performs better too. But maybe the Navy will replace it with yours.

[xor]

Re: Fully Autonomous Sentry w/Target Anticipation

I think this one performs better: http://gizmodo.com/gadgets/tag/robot...you-212241.php

Also there is a thing that is called the MK 15 Phalanx CIWS. I'm pretty sure that performs better too. But maybe the Navy will replace it with yours.

I especially liked their choice of theme music. Auto-turreting to the Pirates of the Carribean Theme song.


xor

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

I think this one performs better: http://gizmodo.com/gadgets/tag/robot...you-212241.php

Also there is a thing that is called the MK 15 Phalanx CIWS. I'm pretty sure that performs better too. But maybe the Navy will replace it with yours.

yea, you fucked up there because those aren't field portable making them a whole different category. already seen them anyway. what you just did was like me saying, "hey, i made a car that drives itself and can negotiate traffic for a fraction of what a regular one costs" and you saying, "but it's not as fast as a veyron." it's good that you brought them up though because it further verifies the lack of competition out there.

[kallahar]

Re: Fully Autonomous Sentry w/Target Anticipation

How come you never entered the contest then? The winning team shot down all the targets in 15.7 seconds with about 90% accuracy...

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

yea, you fucked up there because those aren't field portable making them a whole different category. already seen them anyway. what you just did was like me saying, "hey, i made a car that drives itself and can negotiate traffic for a fraction of what a regular one costs" and you saying, "but it's not as fast as a veyron." it's good that you brought them up though because it further verifies the lack of competition out there.

Yeah I knew you meant but I was just giving you a hard time because of what you actually wrote.

So why is it exactly that you feel like you or what you have built is so incredibly special?
Please be specific, and try not to leave anything out :)

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

How come you never entered the contest then? The winning team shot down all the targets in 15.7 seconds with about 90% accuracy...

because, A; i didn't know about it, B; can't afford to travel there, C; i would have to dumb my project down and retool it to perform an inferior task when it already does something better, more advanced, and more useful.

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

when it already does something better, more advanced, and more useful.

Will you please explain how you have convinced yourself that your creation is useful?

[theprez98]

Re: Fully Autonomous Sentry w/Target Anticipation

I think he wants a contest where he's the only contestant and he automagically wins.

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

Will you please explain how you have convinced yourself that your creation is useful?

when i started on the project in January, whether it was useful or not wasn't even on my mind. but by the time that little demo video i made reached half a million views, i was getting bombarded by people who wanted to use it for paintball scenario games, airsoft scenario games, local and remote property defense, and the all important 'shoot the cats that get into my yard.' (and that's just the civilian side.) so i ask you, compared to that, how in the fuck does shooting still targets, of a singular color, in a controlled environment, with a black backdrop even compare?

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

I think he wants a contest where he's the only contestant and he automagically wins.

i want a contest where moving targets are shot at in an outdoor environment and scored on accuracy. but i guess that would be too hard. gotta keep it simple so there will be enough contestants i guess.

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

when i started on the project in January, whether it was useful or not wasn't even on my mind. but by the time that little demo video i made reached half a million views, i was getting bombarded by people who wanted to use it for paintball scenario games, airsoft scenario games, local and remote property defense, and the all important 'shoot the cats that get into my yard.' (and that's just the civilian side.) so i ask you, compared to that, how in the fuck does shooting still targets, of a singular color, in a controlled environment, with a black backdrop even compare?

I'm sorry, I did not realize how important you were. Nor did I understand the far-reaching contributions to society your suitcase-robot-thing has offered. My friends and competitors have just been having fun with our robots, but now I understand you have been enjoying great commercial success with yours. Hey...Maybe you should take a little of that revenue and start that competition you want so badly! You could even host it in that wooded area outside of the Rivieria. You had better make sure the course it precisely repeatable though or the competitors will start to complain. But that should be easy, you could just make sure some guy (the target) is running around out there the exact same way every round of competition.

If you have some time to spare after that then you should take some of your vastly superior computer vision skills and apply them to some real world problems, like object detection for use on autonomous vehicles (DARPA Grand/Urban Challenge). I think the Stanford guys could use someone like you to show them how to do things.

If you still have time left over I suggest learning to use google. Because there are already better products than yours in production and for sale out in the free market at a similar cost to what you have proposed for yours and of much higher quality. And in case you are wondering...my stuff is not for sale...its a work of love :)

[xor]

Re: Fully Autonomous Sentry w/Target Anticipation

Responding to numerous quotes I'm too lazy to cut and paste sorry


David08;

When requesting expert constructive criticism / advice, data, source code, and hardware diagrams are usually required by experts to give you what you are looking for. Being that you are planning on selling it I don't think that it's in your best interest to make those plans public without a patent or some type of NDA between the parties. Showing us a youtube video really doesn't cut it other than to produce the comments that you apparently don't want to receive.

You could also ask a specific question, like how do I improve the feedback response time of my xxx servo loop. Or why is my xxx current sink showing instability at turn on? You know tech stuff like that. Showing us the video will get comments like oh that's cool. Ask and you shall receive, however asking us to read your mind is pushing it. Come back in 5 - 10 years and maybe we will.

Though robotics isn't my thing I do have a background in Test Engineering. I recommend that if you can get yourself some high-speed film or video recording equipment that it will give you a visual picture of what the device is doing so that you can further improve it.

If you aren't already using them I suggest some tri-axial accelerometers to handle the recoil as well as other non-desired movement. Nothing in the world is static, everything moves all of the time. Static and rigid are the exception.

Of course paint balls don't always travel straight through the air and you mite want to move up to something with a little less air resistance. If you chose bullets please find a non-living target other than yourself. Also, and with all seriousness please don't hurt animals or I'll have to report you to PETA(they are worst that than police).

Again without real material to look over it's very difficult to give you what you are looking for. Hope that helps.

David08, you seem like a really smart guy. Please keep in mind that there is a fine line between confidence(a good thing) and arrogance(a bad thing). Please read the forum rules. As clever as I think I am, when I go to Defcon I'm ALWAYS humbled by the kind and like minded people I find there. I hope you have a chance soon to experience a con and see what I'm talking about. It's sort of like the wild west, there's always someone faster than you.


Good Luck :-)

xor

Actually not to beat a dead auto-turret here. If my offer some fatherly advice:

Dad Like Comments On:

If you are a interested in Weapons Systems Design the fast track to a career doing that is Military(Navy or Airforce), then college at least 4 years. Just like if you are interested in Nuclear Power, first the Navy Nuke Program, then college. Autonomous Warfare probably the Airforce they seem to be getting the most money throw at them today then college.

Dad Like Comments Off

[xor]

Re: Fully Autonomous Sentry w/Target Anticipation

This brings up a very interesting thought.

There are plenty of young deserving minds out there that don't have the resources to travel to Defcon. Being that many people on this list are secure in there careers and have, dare I say extra money; would it be such a terrible idea to maybe start sponsoring some of the next generation? Or do speakers get a free trip already?

xor

[kallahar]

Re: Fully Autonomous Sentry w/Target Anticipation

Also, the defcon crowd isn't all that hardware oriented (at least by numbers). If anyone really wants to start an advanced engineering contest, I highly recommend RoboGames in San Francisco. They definately attract a less hacker and more engineer demographic, and they'll let you run your own contest at their event.