Fully Autonomous Sentry w/Target Anticipation

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

...any idiot can just pay others to do their work for them and make the customer pay for it in the end.

I know what you mean. I was furious when I found out my local grocery store wasn't growing their own produce. They were subcontracting it out to farmers! Profiteering sons of bitches.

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

I know what you mean. I was furious when I found out my local grocery store wasn't growing their own produce. They were subcontracting it out to farmers! Profiteering sons of bitches.

yup.. anyway, i've already tried the machine shop route and found the pricing to be ridiculous. I'm creating a win-win situation for my customers and myself. they win because the price is affordable and i win by selling more volume due to that fact.

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

I am going to omit the sarcasm for this reply.

David08:
It seems apparent that your envisioned product and intended market is different than that of the Zero-Op guys'. Perhaps it could be said that the Zero-Op machine is analogous to a high performance sports car whereas you are striving to create a more affordable automobile. Maybe Zero-Op wants to create products like Enzo Ferrari did and maybe you want to create products like Henry Ford did. Those are both very respectable individuals.

Generally, higher performance requires higher costing materials and methods. Generally, to achieve lower costs at lower volumes sacrifices must be made in terms of quality. Generally, to produce something in high volumes there are usually initial tooling costs which tend to be very high. Generally, those initial tool costs are factored in to the price of the product.

I say let the marketplace judge which product is better. I invite you David08 to "put up or shut up". Put your product up for sale and let the consumer be the judge. This is what the Zero-Op guys have done. If you cannot do the same I will have no choice other than to label you as one who "just talks".

-wsbpress

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

It seems apparent that your envisioned product and intended market is different than that of the Zero-Op guys'. Perhaps it could be said that the Zero-Op machine is analogous to a high performance sports car whereas you are striving to create a more affordable automobile. Maybe Zero-Op wants to create products like Enzo Ferrari did and maybe you want to create products like Henry Ford did. Those are both very respectable individuals.

nope, both target markets are the same. mine already has more features but it costs less to make. the way i see it his is like a ford priced like a Ferrari and mine will be a bmw priced like a ford :)

[kallahar]

Re: Fully Autonomous Sentry w/Target Anticipation

Reliability is a big factor too. For anything over $20 I expect it to be able to stand up to its intended use. For example, that camera enclosure. The zero-op one is designed to be hit hundreds of times with a high velocity paintball. The cheapest way is probably a 1/8" plexiglass (hope that's strong enough, needs product testing) with a bent sheet metal frame. You'll also need a gromet to seal the front in, a mounting rail for the camera, gromets for the wires, and a mounting post on the outside.

All those costs add up...

Also, I could easily see a paintball field buying two of these for a field. Let the teams rent them for $20/game and they can place them wherever they want on the field. They're going to want continuous reliable use for their money.

I think it's clear who has the real world experience here. I agree with wsbpress, put up or shut up :)

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

nope, both target markets are the same. mine already has more features but it costs less to make. the way i see it his is like a ford priced like a Ferrari and mine will be a bmw priced like a ford :)

Ok great. You shouldn't have any trouble making tons of money on it then. When exactly will these be available? Why haven't you started selling these mechanical gems already? You've got it all figured out already so what's the delay?

Though this has been all very entertaining I am going to give up on you now. You have your work cut out for you and you probably shouldn't be wasting your time posting nonsense.

Good luck with your endeavor. I hope you are able to do at least some of the things you have been talking about.

-wsbpress

[wsbpress]

Re: Fully Autonomous Sentry w/Target Anticipation

I feel like I've been on the set of Jerry Springer or something so I need to post something positive :)

Thanks Kallahar for your time and effort in hosting the defconbots competition for these last several years. And to those who are unaware there is currently an open discussion about what form next year's competition should take. Instead of messing with with Davey I think I will try and help out over there with the brainstorming effort.

-wsbpress

[xor]

Re: Fully Autonomous Sentry w/Target Anticipation

David08, you say it takes you 4 days to make your product using your hacksaw and hot glue or whatever (no machine shops for you). Lets say (hypothetically) that your bot costs you $500 (which you say it does) in materials and you sell it for $750. Your margin is $250 and assuming 8-hr work days you've spent 32 hours of your time building this bot. So you are making about $7.80/hr but that isn't including any distribution or marketing costs. And if you are paying that high school kid minimum wage ($6.79 in your state) then your cut is only about $1.00/hr. So, after 3 hours of QA'ing your high school friend's work you could afford to take a break and get a Happy Meal. Make sure your friend drives though because its going to take you the equivalent of 4 hours of work for you to purchase one gallon of gas.

-wsbpress

Isn't that what Chinese people are for? I know terrible; just saying what I know everyone is thinking.

xor

[Greyhatter]

Re: Fully Autonomous Sentry w/Target Anticipation

Isn't that what Chinese people are for? I know terrible; just saying what I know everyone is thinking.

xor

Child labor! Oh Chinese labor, whew I though you were talking about slave labor or something...

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

Ok great. You shouldn't have any trouble making tons of money on it then. When exactly will these be available? Why haven't you started selling these mechanical gems already?

because I'm too much of a perfectionist. i've been emailed by plenty of people who tell me to quit tweaking it and just sell them already. i don't want to make the same mistake cfinger made..

[miked]

Re: Fully Autonomous Sentry w/Target Anticipation

davido8 - Your ideas are intriguing to me and I wish to subscribe to your newsletter

[Wesson357]

Re: Fully Autonomous Sentry w/Target Anticipation

Thanks :)

Yeah, I really understand the cost of producing something now, as I'm sure you've seen w/ the kits you built. I don't think we want to mention too much of our production costs, but I think we've done a good job finding a cheap route to go. Its priced pretty competitively in my mind, and we're kind of counting on hardware costs dropping at higher quantity.

A lot of cost comes up when your goal is to make everything waterproof and durable. Any custom work, especially machining, costs a good deal up front. A good example of that is our camera enclosure. Look up any commercial waterproof camera enclosure, and you'll pay anywhere from $300 to $600. Not to mention there isn't a standard for paintball-proofing :)

Your video is quite interesting. One question, it seems all of your targets in the video had extremely visible solid color shirts that contrasted well with the background. How does the system perform on people who are camouflaged a bit more?

[David08]

Re: Fully Autonomous Sentry w/Target Anticipation

davido8 - Your ideas are intriguing to me and I wish to subscribe to your newsletter

just visit the site. it's updated often.

[cfinger]

Re: Fully Autonomous Sentry w/Target Anticipation

Your video is quite interesting. One question, it seems all of your targets in the video had extremely visible solid color shirts that contrasted well with the background. How does the system perform on people who are camouflaged a bit more?

Yeah, so originally we just used motion to track targets. There are 2 reasons to use color AND motion-

1- Teams. Red vs. Blue. So it can be part of the game to advance up to the turret, switch its team, and turn it towards your enemy.

2- If it works solely off motion, what happens when the wind blows, trees start to move, grass moves, everything is moving! So with our software (you can actually see this in the video) even when all that tall grass is moving in the wind, the system can ignore that and focus on one target. Same thing w/ color - you can see red train cars in our video, which aren't being shot at.

Computer Vision people know the biggest, and arguably the hardest, part of the system is making it work in all situations. We have simulation software set up so that we can record hours of video in all different situations, come back to our shop and find the best parameters that will work universally. THAT is something that takes time and a lot of patience- the more you tailor it to one situation, the more you screw up the other situations.

Now the other reason to use color, is to have atleast some wiggle room in the liability department. There is less of a chance of hitting an innocent bystander- still a chance, but less. Just don't wear red or blue :)

[kallahar]

Re: Fully Autonomous Sentry w/Target Anticipation

cfinger - is your camera mounted to the barrel or is it fixed? If fixed, how are you determining the range to the target to figure out the correct angle? Or do you assume it's a level field? I know a lot of teams had trouble calibrating the firing position to where the camera was looking.

Kallahar