PDA

View Full Version : The future of Mech Warfare



RobotAtlas
11-02-2010, 07:00 PM
The current state of Mech Warfare is roughly an equivalent of a First Person Shooter (FPS) video game, but in the real world.
It’s in the early stages and things don’t quite work, but they will in a few short years.

What’s next then?

I’m a big fan of StarCraft and I have a dream today.

In the future we will have a version of Real Time Strategy (RTS) game, but in the real world, with real robots. Let's call it Real Robot Startegy (RRS or R2S)

What’s the main difference between the two styles?

In a FPS game you are micro-managing your robot by issuing lover-level commands:
Turn right. Go. Stop. Shoot. Fall (well, that just happens by itself :-).

In R2S you will be able to issue higher-level commands to your robot:
Go there. Do this. Find that. Bring it back. Repeat until told otherwise. Go home when attacked.

When is R2S going to happen?
I don’t know, but the future has to be dreamed up before it can become a reality.

Slugman
11-02-2010, 08:05 PM
Just my .02 - Given the cost of these things, The single bot - first person shooter situation will last for some time - You will see more advances & complexity, but only on single bots, as no-one can afford to field multiple bots. Given a greater payload capacity, you might see more targeting plates with varied simulated damage, but overall you will just be seeing greater speeds/mobility, along with more autonomous bots - Either auto movement, or more likely auto-targeting. As the number of people creating these increases, you will get an opportunity to create larger battlefields & teams, but it would still be First Person control. BTW - I think things are working now. I played a Mechwarrior computer game years ago that had a damage system only marginally more complicated than the one used in this real Mech Warfare, & the aiming system was the same - Point & shoot - No auto-targeting back in the 80's! I think the big change you will see over the next 2 yrs is a greater use of IMUs for better speed & stability. 10-15 targeting plates with different damage types/effects depending on location would be nice - It was always fun to target the legs of an opposing mech to stop them, then hit the guns.... Would require a bit of extra programming to shut down one leg of a biped based on damage, but would be quite funny to see. :cool:

lnxfergy
11-02-2010, 08:44 PM
Would require a bit of extra programming to shut down one leg of a biped based on damage, but would be quite funny to see. :cool:

Well, 99.999% of currently existing robot bipeds would basically fall over if you shut a leg down......

-Fergs

Slugman
11-03-2010, 12:14 AM
Yes, but if there was something in the Warfare rules like "If the impact sensors on any single leg of a robot detect a total of 10 hits or more, then that leg must become rigid & cannot move from that moment on, for the duration of the session." then you would find everyone programming in alternative gaits to ensure their bot stays in the game. You gotta admit, it would be pretty amusing to see a quad dragging themselves around the arena with one leg in a close, multi-bot combat, or a biped using one arm as a prop. Now THAT sort of footage would make the news. :happy: Would also need some points balancing to make sure quads (& Hexes in open) weren't overpowered compared to bipeds. My main point was to give an idea of my view of the direction that Mech Warfare might take, without getting into the details too much - The cost of so many small plates could be an issue, let alone the complexity of making such a system work in real-time, but then again, that's why it is in the future, I hope. (Not that I will ever be able to justify flying 1/2-way around the world to actually attend one of these events - *sigh*)

master_of_robots
11-03-2010, 08:41 AM
...by issuing lover-level commands...

I like "lover-level commands" :veryhappy:

But then, I think we would be talking about pleasure bots rather than kill bots. :rolleyes:

darkback2
11-03-2010, 10:16 AM
I'm not so sure the swarmbots, teambots is so out of the question. Consider wheel based mechs using one computer controlling several mechs via xbee. Its impossible considering the cost per bot/difficulty would be dramatically lowered. four bots on a team is not impossible especially if you have uni's getting involved.

I'm more for alternate campanes/team warefare.

As for adding realism, I could see a time out being implemented really easily. Whenever a bot registers a hit, it stops taking commands for a couple of seconds, or returns to a base position for a few seconds.

Another idea I had was to have someone with an airsoft gun sit behind the pilot. Every time their bot gets hit...

:)

Connor
11-03-2010, 03:41 PM
I'm not so sure the swarmbots, teambots is so out of the question. Consider wheel based mechs using one computer controlling several mechs via xbee. Its impossible considering the cost per bot/difficulty would be dramatically lowered. four bots on a team is not impossible especially if you have uni's getting involved.

I'm more for alternate campanes/team warefare.

As for adding realism, I could see a time out being implemented really easily. Whenever a bot registers a hit, it stops taking commands for a couple of seconds, or returns to a base position for a few seconds.

Another idea I had was to have someone with an airsoft gun sit behind the pilot. Every time their bot gets hit...

:)

I wish that rule was implemented last year during our match DB. :) You would have been one sore dude. :)

DresnerRobotics
11-03-2010, 04:25 PM
No offense, but I think people should probably focus on actually building reliable mechs before discussing where the competition will or won't head in 5 years. We need a larger player base for it to evolve into a more advanced war game. I'm also going to be far more likely to take feedback from veterans of the competition.

RobotAtlas
11-03-2010, 07:22 PM
No offense, but I think people should probably focus on actually building reliable mechs before discussing where the competition will or won't head in 5 years.

I'm glad you mentioned it, Tybes.
I believe every single person needs to learn how to do RC befofe he or she can attempt to do autonomy.

And I'm going to put my money where my mouth is by buying an Interbotix hexapod kit (unless it's grossly overpriced, which is unlikely).

The question I do have undecided: can one go straight to a walker or should one start with a wheel-based robot?

P.S. > in 5 years.
Andrew, I didn't know you are an optimist.

sthmck
11-04-2010, 08:57 AM
I'm glad you mentioned it, Tybes.
I believe every single person needs to learn how to do RC befofe he or she can attempt to do autonomy.

And I'm going to put my money where my mouth is by buying an Interbotix hexapod kit (unless it's grossly overpriced, which is unlikely).

The question I do have undecided: can one go straight to a walker or should one start with a wheel-based robot?

P.S. > in 5 years.
Andrew, I didn't know you are an optimist.

I don't know if tybs is actually referring to building an autonomous versus tele-operated, as much as he is referring to building a mech that actually functions like a competitive mech should. If you have any successful experience in getting a rover to be autonomous then it really shouldn't matter if you want to make your mech RC or autonomous.

Whether you have built a rover before shouldn't play as much of a factor in your decision for which type of mech you build. What you should really consider is the skills that you have and your ability to learn.

As for purchasing an interbotix kit. If you are planning on going this route most of the hard work for building a mech is already done for you. Tybs has created a very solid platform to start with, and with NUKE fergs has basically taken care of 80% of the programming you are going to need to do. The real question is do you want to put the money down up front for your mech and pay a little more for a solid kit, or do you want to do everything yourself and pay for mistakes you will end up making during your build?

UncleBob
11-04-2010, 09:19 AM
I think what tybs mean is instead of trying to predict the future of the sport / hobby, we can do something at the present e.g. building a robot of some sort. This way we have more knowledge and we can predict if we want to or create the future perhaps.

Having said that having a few futurist might also be helpful e.g. Arthur C Clarke wrote scientific novel about space station before it is possible. This gives other people the imagination to create.

sthmck
11-04-2010, 09:33 AM
I think what tybs mean is instead of trying to predict the future of the sport / hobby, we can do something at the present e.g. building a robot of some sort. This way we have more knowledge and we can predict if we want to or create the future perhaps.

Having said that having a few futurist might also be helpful e.g. Arthur C Clarke wrote scientific novel about space station before it is possible. This gives other people the imagination to create.

Couldn't agree with you more. Looking at the future is great, but due to the fact that this is a COMPETITION the most changes will come from people actually building. The sport will evolve and based on the number of competitors and the overall innovation brought to the table by capable builders.

RobotAtlas
11-04-2010, 02:50 PM
Great post sthmck (BTW, I have very hard time pronouncing that name. JHow do you pronounce it?).


What you should really consider is the skills that you have and your ability to learn.

The real question is do you want to put the money down up front for your mech and pay a little more for a solid kit, or do you want to do everything yourself and pay for mistakes you will end up making during your build?

These two questions address past (skills you already have) and present (what do you feel like learning right now).
Three additional questions could be asked to ensure future satisfaction in this hobby.
I wanted to post those questions here, but that would be a long post, so I put it into a separate blog:
http://forums.trossenrobotics.com/blog.php?b=273

I do want to add to “kit/no kit” discussion here:

If I build my own robot without using Interbotix kit, it’s going to take me some time before I get it to the point where Interbotix is today.
How long might it take? Tybes started messing with Bioloids about 2.5(?) years ago, plus he had experience with hobby servos/robots before that.
On the other hand, since I’m not pioneering here, I can use Tybes (and others here) experience to shorten my learning curve.
So, if I’m great and spend enough time, I might be able to get where Interbotix is today in 6-9 months.
By that time Interbotix will move forward too. So now I’m playing a catch up game. But I’m also learning a lot.

Alternatively, if I get Interbotix kit as soon as it gets available and spend time studying/playing with it, I will be able to improve on it.
What do I want to improve though?

That’s where my new blog post comes into play.

Following that logic, I decided that I’m primarily a beta-tester aspiring to be a researcher. I want to improve on already existing robots.
I’m a software engineer with electrical engineering background. That narrows down what type of things I want to improve.

elaughlin
11-04-2010, 04:59 PM
I just thought I could add alittle about what you said RobotNV...




If I build my own robot without using Interbotix kit, it’s going to take me some time before I get it to the point where Interbotix is today.
How long might it take? Tybes started messing with Bioloids about 2.5(?) years ago, plus he had experience with hobby servos/robots before that.
On the other hand, since I’m not pioneering here, I can use Tybes (and others here) experience to shorten my learning curve.
So, if I’m great and spend enough time, I might be able to get where Interbotix is today in 6-9 months.
By that time Interbotix will move forward too. So now I’m playing a catch up game. But I’m also learning a lot.



You may notice that my Mech has a lot of very similar parts to the Interbotix. It is because I started off at the same point as you I think. I had no robotics experience at all, I just happened to come across an article in a magazine I had never seen before when waiting around in a Borders bookstore. It was an article about Mech Warfare. I was struck right away. I did research, looked around the web, and happened to make it to this website where I met these Roboteers.

I had the same questions that you had posted before, because like most people, I want to make a sweet ass looking biped like the video games portray. I talked to Tyberius for a couple of weeks before I decided on getting the Bioloid Kit. Then I happened to see small pictures of the Interbotix that Tyberius was playing around with at last years Robogames. I was like "I want to make it like that!"

So Tybs helped me out, and showed me the legs, and I went from there. I am pretty much using the Interbotix kit before it came out, but am also personalizing it to what I want out of it.

For me then, the timeline started about 8 months ago in March when I got my Bioloid kit. Then at the end of April I started the work on designing the pieces off the Interbotix model. Therefore, I would say by going with the kit, you are saving yourself about 5 months of work (which I worked on mainly 4-5 days a week, starting from no ability with any program or experience) and probably going to save about $300-$500 dollars, I would guess by buying a kit from Trossen. I think I am past the $2000 mark already. When they group stuff together, they really price it sell. Plus, I think the kit is still easily adaptable to be changed and still be a very unique bot compared to other kits, if some time is put in.

That's just my experience so far with working with Robots. And I am stuck in this now. I don't plan on ever stopping my roboting ( I make up my own words, maybe it will catch on).

Stobs
11-05-2010, 12:05 AM
@RobotNV: here's a link for the InterbotiX hexapods (http://www.trossenrobotics.com/lynxmotion-phoenix-hexapod.aspx) (found by searching for "hexapod" rather than "interbotix'); I guessed at some of the pricing but it seems to me the basic kit is within about $200 of purchasing the parts individually at retail through Trossen or other sources - not counting the Lynxmotion Phoenix no-servo/no-electronics kit (http://www.lynxmotion.com/p-650-phoenix-3dof-hexapod-no-servos-no-electronics-black.aspx), which retails for ~$250. I'm not going to say yay or nay about kits ("mostly" because I haven't made up my mind about using one or not), but I'd easily say that from a retail perspective their definitely not overpriced, grossly or otherwise.

Regards,
Stobs/Paul

darkback2
11-05-2010, 12:34 AM
My first walker was a kit. When I was ready to move on, I gave most of it away. Kits are a great place to start, especially when you are new at something. really, very few people actually contribute anything new, myself included. most of us are simply living closer and closer to the edge of contribution.

There does come a time though when talking isn't going to work anymore, and doing is what matters. You will always be playing catch up, so you might as well take the plunge...The water is fine,

elaughlin
11-05-2010, 07:42 AM
@RobotNV: here's a link for the InterbotiX hexapods (http://www.trossenrobotics.com/lynxmotion-phoenix-hexapod.aspx) (found by searching for "hexapod" rather than "interbotix'); I guessed at some of the pricing but it seems to me the basic kit is within about $200 of purchasing the parts individually at retail through Trossen or other sources - not counting the Lynxmotion Phoenix no-servo/no-electronics kit (http://www.lynxmotion.com/p-650-phoenix-3dof-hexapod-no-servos-no-electronics-black.aspx), which retails for ~$250. I'm not going to say yay or nay about kits ("mostly" because I haven't made up my mind about using one or not), but I'd easily say that from a retail perspective their definitely not overpriced, grossly or otherwise.

Regards,
Stobs/Paul

That link isn't actually the Interbotix Hexapod. The one you posted is the Lynxmotion Phoenix. Interbotix Hexapod is a creation made by the team at Trossen. That also uses hobby servos as you can see and not AX's. I am guessing that Trossen has to order the Phoenix from Lynxmotion to sell at TR, and therefore is just selling it at the price they buy for, or maybe some profit.

I don't think the Interbotix Hex is up for sale as of yet. It is just something talked about, that is in production to be sold soon. The pieces for it are custom created by the team, and therefore they mass produce it themselves for a much lower cost, and therefore the savings of the Interbotix will probably passed on to us, the customers of TR. Here is the link to what the Interbotix (http://www.trossenrobotics.com/productdocs/assemblyguides/phantomx-quadropod.html) looks like.

UncleBob
11-05-2010, 09:46 AM
Before I bought my first bioloid I have experience with rc helicopter. But I have no idea about robot. I wouldn't know how to put it together and what else is involve. I think I would need to learn off an existing robot in order to learn enough experience to build a robot from scratch.

So for me I would prefer to buy a kit to learn the basic first. I would also need to figure out how to program it. Otherwise It won't be able to move even if I manage to put it together.

Having a kit gives me instant satisfaction and encouragement. On the other hand building from scratch would be boring for many months.

RobotAtlas
11-05-2010, 10:52 AM
really, very few people actually contribute anything new, myself included. most of us are simply living closer and closer to the edge of contribution.

darkback2, you are being too modest here. I bet I can find many new things you contributed.
“New” is a tricky thing. What’s not new to you is new to somebody just starting.
Also “reminded” is often as good as “new”.
In your last post you already contributed encouragement to stop talking and start doing something.

On a related note, people don’t need to wait until they reach certain level of knowledge before they can start contributing.
To give a real-life example:
You learn something by doing a tutorial and you find an error there.
Go ahead and contact the creator of that tutorial and let them know.
This way the next guy will have an easier time doing this tutorial.

If you just look for ways of giving back, you can always find something.

Stobs
11-05-2010, 04:49 PM
That link isn't actually the Interbotix Hexapod. The one you posted is the Lynxmotion Phoenix. Interbotix Hexapod is a creation made by the team at Trossen. That also uses hobby servos as you can see and not AX's. I am guessing that Trossen has to order the Phoenix from Lynxmotion to sell at TR, and therefore is just selling it at the price they buy for, or maybe some profit.

I don't think the Interbotix Hex is up for sale as of yet. It is just something talked about, that is in production to be sold soon. The pieces for it are custom created by the team, and therefore they mass produce it themselves for a much lower cost, and therefore the savings of the Interbotix will probably passed on to us, the customers of TR. Here is the link to what the Interbotix (http://www.trossenrobotics.com/productdocs/assemblyguides/phantomx-quadropod.html) looks like.

[Edit: I stand corrected; didn't realize that Trossen would be offering competing hexapod kits - my bad!]

Stobs/Paul

UncleBob
11-05-2010, 05:15 PM
I notice a lot of hexapod and quadpod lately. Is this getting popular?

Just asking
11-05-2010, 07:11 PM
Yes, mostly because hexapods have more of an advantage over bipeds. Bipeds would just fall down if they stepped on a BB, rock, and just about anything bigger than a BB while hexapods a quads don't. Hexapods also have more speed. For example, Giger is a great robot, but it lacked speed. A cheap hexapod defeated it last year, no offence Tyberius.:)

Upgrayd
11-05-2010, 07:31 PM
cheap!?

Just asking
11-05-2010, 08:06 PM
Well sometimes depending on the materials. Remember, this is in robotical terms. For example,
you spend 800$. It might sound a lot but you really saved an extra 100$-200$. I could use
that money to buy myself some extra gear for my mech.

sthmck
11-05-2010, 09:11 PM
hexapod?

sthmck
11-05-2010, 09:24 PM
Honestly the cost of a high quality mech thats capable of winning the competition can't be calculated in parts cost alone. You should probably consider the amount of time spent in research and development. Some people choose to invest in skills required to build a good robot rather than dumping large sums of money into something that is essentially useless. I say this with no disrespect to Tybs, I know he put as much if not more work into his robots as anyone else that competes, but you have to consider what upgrayd did. The cost of building his mech was in the time he spent to take what he had and produce a top contender. If you consider the skill set he has and the amount of time put into it you will realize that his work has quite a bit of value in and of itself.

DresnerRobotics
11-05-2010, 11:46 PM
I'm recovering from a nasty flu I picked up in Korea, but wanted to chime in on a few things in this thread. I'll be back in full swing next week.

First of all, I believe credit must be given where credit is due. Keep in mind that while yes, the Trossen team is bringing the Interbotix hexapod and quad kits to the market (and even a Mech Warfare specific Quad Starter Kit Shhhh! :P ), they're a reality due to lnxfergy for his ArbotiX/NUKE solution, which are the core of the kits and what makes them so easy to put together. He also single handedly designed the scoring system, and has been an integral part of making Mech Warfare happen. He's put just as much effort (if not more) than I have into Mech Warfare, so make sure you give him a nice +rep sometime when you get a chance, he's a fricken rockstar.

In terms of kits vs scratch built. Honestly, some people take great pride in going scratch built, some people just roll with kits (modified or otherwise) and neither is right or wrong. Do whatever works best for you and your timeline, but there's no shame in using a kit. One of the things that will be great about these upcoming kits is that we're combining some pretty awesome electronics (the arbotix gear), with some custom CAD designed laser cut parts from a pretty expensive laser CNC, with AX-12s to put together a kit that often times would be out of reach for your average hobbyist. The intent here is to get more people involved in Mech Warfare. If we can literally link them to a starter kit to get a head start, that opens up the competition considerably, and a larger player base means a faster growth as a competition. I'd love to see the 'bar' of performance set at an AX-12+ quadrupod based kit, because that means better piloting, control, gaits, and aiming/firing systems are going to make or break you. It shouldn't be a struggle just to get your bot walking.

And no offense taken Seth- I had long said "I'm totally calling it, sub $1000 quad is going to kick Giger's ass". I built an RX-64 quad using an Arbotix + Bridge that would've made a far more effective Mech Warfare platform, yet never used it as such because I felt Giger was more of a challenge. I will say one thing, after having built 2 smaller bipeds in the last month, big humanoids (>5kg) are REALLY, REALLY, hard to get walking well. Dumping money into this competition really doesn't get you very far (nor was that ever my intent with Giger, I just wanted a big 'show bot'), Upgrayd's quad or Fergs' Issy1/2 are perfect examples of this. Superior engineering, control solutions, and time spent practicing will ALWAYS trump dollars spent in my opinion. Upgrayd beat me handily last year because he had a very well polished Mech with fantastic control and agility, and he was a great pilot to boot.

master_of_robots
11-17-2010, 09:16 AM
this is in robotical terms

"Robotical" LOVE IT! I'm going to have to use this new term in a sentence today.

kensbey
12-07-2010, 08:09 PM
going back to the original thread topic, (r2s) i think that there will be an intermediate stage concerned with automation. i would love so see a mech warfare competition that is 100% autonomous and onboard processed. everyone gets their bots, goes to some location, programs in the combat boundaries, then stands back and watches their creation flourish/flounder.

of course, as mentioned, a larger participation is fundamental in the development of the sport.

YAY THE FUTURE

darkback2
12-08-2010, 12:36 AM
I have been playing with color tracking for Hikari. I found a way to click on a part of the video image and have the computer track the point where you clicked. I don't have it 100 percent yet, but is seams like you could use a similar solution to "lock on" to a target, and continue shooting. The same place regardless of the fact that you are walking around avoiding being shot yourself.

DB

RobotAtlas
12-08-2010, 09:25 AM
but is seams like you could use a similar solution to "lock on" to a target, and continue shooting. The same place regardless of the fact that you are walking around avoiding being shot yourself.
Now we are talking. :)
If a mech had SLAM capability, you could just click on the map to set waypoints and then just concentrate on shooting. If mech is moving all the time, shooting becomes problematic. Then this "lock on" feature would be very handy.

lnxfergy
12-08-2010, 09:31 AM
Now we are talking. :)
If a mech had SLAM capability, you could just click on the map to set waypoints and then just concentrate on shooting. If mech is moving all the time, shooting becomes problematic. Then this "lock on" feature would be very handy.

I believe I have said it before -- but -- there is no need for SLAM in mech warfare -- we don't need to map the environment -- it's known ahead of time, it doesn't change either. Basically, you're looking for localization within the environment against a known map.

-Fergs

RobotAtlas
12-08-2010, 09:52 AM
Today we know map. This thread is about future.

lnxfergy
12-08-2010, 10:17 AM
Today we know map. This thread is about future.

Trust me -- as soon as you've helped set up/tear down you'll understand that the map really isn't going to change (even if it does, it wouldn't happen more than once a day).

-Fergs

sthmck
12-08-2010, 11:27 AM
Mech SLAM: System Localization And Movement... I made everyone happy... not. Seriously though people who have never made a mech really tend to have no clue what they are talking about when referencing the future. People tend to make a lot of assumptions about the future based on stuff they have little experience with, when in reality what they are actually calling the future is most likely something that fergs will end up cooking up in his spare time. run-on sentence yay!