PDA

View Full Version : Initial updates from Robogames...



jes1510
06-13-2009, 11:47 AM
We needz 'em. What are the first impressions from the competitors? How are the mechs holding up? Is Giger a smoking pile of scrap, a victim of Charlie's Rocket powered awesomeness? Has Shadow scout been able to walk around without falling over? Come on guys keep us poor slobs that couldn't afford to go informed!

Testing2000
06-13-2009, 12:13 PM
I'll give an update tonight. Adrenalynn sounded slightly, I repeat slightly frazzled when I spoke with her this morning. That means that she is on the problem! This town is messed up. the streets were layed out by madmen.
.

lnxfergy
06-13-2009, 12:43 PM
We needz 'em. What are the first impressions from the competitors? How are the mechs holding up? Is Giger a smoking pile of scrap, a victim of Charlie's Rocket powered awesomeness? Has Shadow scout been able to walk around without falling over? Come on guys keep us poor slobs that couldn't afford to go informed!

Check Issy's thread... I posted about his progress... or lack thereof....

Giger has been sitting around with his head in his hands most of the time...

-Fergs

engineer
06-13-2009, 03:10 PM
us huminiods did demo from 3:00 til I left at 7:00 I think RN-! messed another servo I'm going to wait til my first round to see if I need to change it nothin like pressure to get RN-! apart and back together between matches. Table space is real tight, what I set up thursday is all messd up. hoping to kick some bot

lnxfergy
06-14-2009, 12:08 AM
Mech coverage can be found here: http://forums.trossenrobotics.com/showpost.php?p=31316&postcount=34

-Fergs

darkback2
06-14-2009, 02:42 AM
Ok...everyone is dogging me because my quad fell over...But a damn leg fell off! Ok...so not all the way off, but I had a screw back out of a 5990 hip servo on squidword, and he fell over. So I've been expelled from mechwarfare for life...

then again, who else can shoot rockets?

thats what I thought!

DB

Adrenalynn
06-14-2009, 04:29 AM
LandShark attempted suicide by jamming a screwdriver into his own heart. Thanks to Andrew and Fergs - he went to the top ofthe donor list... Blown-up XBee anyone?

Oh - and XBee Series 2 doesn't talk to XBee Series 1...

'Course, he's not a mech, but people do seem fascinated by the arm controller...

robologist
06-14-2009, 08:32 AM
All great info, sounds like a lot of fun. Thank you all for the updates, good to read!

nagmier
06-14-2009, 09:13 AM
You are all bastards!

LUCKY... but bastards none the less :D

Great to see everyone is having a good time, I hope we have some people taking video as well! I would love to see the wars in action! Good luck today Ladies and Gentlemen!

mannyr7
06-14-2009, 11:21 AM
OMG, I hope somebody has video on that! Shoot, I left too early. Be back sunday morning. Clyde was moving and shooting late Friday night, just had some tweaking to do on his gaits.
Then Saturday afternoon, my Bluetooth glitched, Clyde drove his aluminum head into his own unprotected SSC-32 and I was done for after that. Hope to borrow a spare today and get into the fight.

elios
06-14-2009, 10:42 PM
hey guys, i might have missed it, but is there any video footage to us eager fans...?

cheers

Adrenalynn
06-15-2009, 10:30 AM
The event ended late last night. No. There's no video footage. It takes time to edit half a dozen cameras for hours of video into a single little youtube clip, ya know?

mannyr7
06-15-2009, 11:54 AM
Can't wait to see! I had my nose in a soldering iron all weekend. Many thanks to Andrew, Trossen Robotics and whole Robogames crew for putting on a great event. Me and the kids had a blast and it was a pleasure meeting everybody in person. It's a shame Clyde didn't get working a couple hours earlier, but it was a learning experience and I got a headstart on next year's games. Look out in year II!

ooops
06-15-2009, 01:26 PM
Here are some pictures.

ooops
06-15-2009, 01:28 PM
And some more:

gdubb2
06-15-2009, 09:03 PM
I had a wonderful time.. It was great to get to meet everyone. Thanks to Alex, Andrew and all the Trossen crew. A big thanks to Fergs, Connor, and Jodie for the help.

Looking forward to next year.. I solved a lot of problems in my head during the drive back to Oregon.

Gary

sthmck
06-15-2009, 09:51 PM
Its great to hear that everyone had fun. I cant wait to hear all the stories as well as all of the ideas for improvements that I am sure everyone has. I am already looking forward to being able to go next year.

DresnerRobotics
06-15-2009, 11:41 PM
So. Tired.

Will post more later...

lnxfergy
06-16-2009, 01:05 AM
In that last photo Ooops posted, that lazer dot on Matt's forehead would be from Issy....

-Fergs

elios
06-16-2009, 01:59 AM
The event ended late last night. No. There's no video footage. It takes time to edit half a dozen cameras for hours of video into a single little youtube clip, ya know?

lol why dont you just upload some raw footage as a teaser....! lol!

This is INCREDIBLE!!!!!!!!!!!!!!!!!!!!!!!!!!!! Wish i could have attended. im still saving, and i have some work on the go that might get me somewhere close to my goal. How sucessful were the mechs, like, what was the preferred mobility of choice, quads, bips etc?

This is very interesting and a great learning curve....

~Elios

Adrenalynn
06-16-2009, 02:20 AM
1. The footage belongs to Tybs
2. "raw footage" - I fear you may not understand what that means. 500MB/min/camera = 2GB/min, average match about ten minutes = 20GB for a match. How exactly do you propose I move that around?

ScuD
06-16-2009, 03:34 AM
Shovel and wheelbarrel?
:p

Incidentally, I'm wondering how Tyb's biped did as well.

Adrenalynn
06-16-2009, 03:45 AM
It didn't. That's about all I can say. Other than it taking a header off a table. ;)

librab103
06-16-2009, 08:10 AM
lol.....it looks like bipeds are out and quads are in? Who would of thunk?

Orac
06-16-2009, 11:56 AM
You could always strap two Bipeds together :P

lnxfergy
06-16-2009, 11:58 AM
Bipeds are just gonna take more time and engineering, keeping the weight down, and keeping it well balanced is very important.... they will also cost significantly more than a quad. That said, the bipeds will be far more awesome when walking around gunning at each other... I plan to build a biped brother for Issy by next year.

-fergs

tom_chang79
06-16-2009, 02:13 PM
LandShark attempted suicide by jamming a screwdriver into his own heart.

Did anyone get a video of this??? I want to see this when I get home!!!

Adrenalynn, did you compete in Mech War?

Adrenalynn
06-16-2009, 02:40 PM
I didn't compete - I just did the video stuff.

John from SoR got video of the suicide attempt, I believe. I've gotta drop him a line and see if he has it.

ooops
06-16-2009, 03:20 PM
You know having seen Land-Shark crusing around the Mech Wars arena, maybe tracs should be considered for next year. Just double the hit sensors on them, and keep the compition trac vs trac.

mannyr7
06-16-2009, 04:35 PM
Enjoy!



http://www.youtube.com/watch?v=bGCy2Oh5iBQ

lnxfergy
06-16-2009, 04:49 PM
YouTube - RoboGames 2009 Mech Warfare

I love that they caught me just repeatedly saying "My gun is all frigged up"

-Fergs

ooops
06-17-2009, 08:21 AM
Wow, Tyb looks so Professional:)

lnxfergy
06-17-2009, 03:39 PM
Another onboard view, this time Issy vs. The Intern

YouTube - Issy vs. Intern

-Fergs

nagmier
06-21-2009, 10:06 PM
This is very good news... maybe all my grumbling about bipeds is unfounded... I'm just hoping we can see all our mechs speed up a little bit, if only for the audience's sake.

-Fergs

While I wasn't there I would say that with some of the ideas tossed around in a few other threads, Bot cam views, pilot views, scoreboard and better access (maybe viewports on opposing sides and pilots on opposing sides but offset from the crowd) would certainly help tho I do see your point tho.

Adrenalynn
06-21-2009, 11:47 PM
>> "Bot cam views, pilot views, scoreboard"

How would those help bipeds walk? Oh - and we also had all those. :)

lnxfergy
06-22-2009, 07:06 AM
>> "Bot cam views, pilot views, scoreboard"

How would those help bipeds walk? Oh - and we also had all those. :)

He's sayng those would help the audience not be bored when a biped walks less than an inch per second...

-Fergs

Adrenalynn
06-22-2009, 12:27 PM
Ahh, apologies. Understood.

mannyr7
06-22-2009, 02:42 PM
It would be great to see more reliability and quicker action, but I'd hate to see this turn into another battlebots type event, which are just souped up r/c cars with weapons. Great for the fans, but how is this robotics? To me, mech warfare was about 40-80 ton lumbering behemoths using strategy to outmanuever and outgun your opponent. On paper, Mech creation was a major exercise in compromise. "I can't mount 6 PPC's if I still want to move 100ft per round." We've already seen some of this in action. This year's event was a great start. I'd like for the rules to adapt, if needed, to maintain this focus. In the Mechwarrior/Battletech universe, quads were very uncommon. Not sure why, they make more sense in the real world. But, if it's real world realism we're striving for, the arena floor is flat and paved. Just throw out walkers altogether and make lightening fast wheeled rovers loaded to the gills with Defenders pointing in every direction.

Just my two cents. -Manny

Adrenalynn
06-22-2009, 02:51 PM
I see your point - but we do have to exist in the real world. And in the real world, I didn't see "a lot" of bipeds actually able to compete this year. And that's not a slight to the bip builders - it's a really tough problem!

Like it or not, it's the audience that ultimately pays the bills. If there are only a handful of people being entertained by the event come year two or three, I suspect RG isn't going to be too excited about sacrificing all that real estate. It has to draw a crowd. RG is about the door fees, and to collect the door fees, they have to deliver mass-appeal entertainment.

Personally, I appreciate that Dave and Simone were willing to take a flyer and permit us such a huge amount of floor space for a first-year, totally unproven, event. I think we need to give some thought to how we provide them maximum ROI on that.

lnxfergy
06-22-2009, 02:59 PM
It would be great to see more reliability and quicker action, but I'd hate to see this turn into another battlebots type event, which are just souped up r/c cars with weapons. Great for the fans, but how is this robotics? To me, mech warfare was about 40-80 ton lumbering behemoths using strategy to outmanuever and outgun your opponent. On paper, Mech creation was a major exercise in compromise. "I can't mount 6 PPC's if I still want to move 100ft per round." We've already seen some of this in action. This year's event was a great start. I'd like for the rules to adapt, if needed, to maintain this focus. In the Mechwarrior/Battletech universe, quads were very uncommon. Not sure why, they make more sense in the real world. But, if it's real world realism we're striving for, the arena floor is flat and paved. Just throw out walkers altogether and make lightening fast wheeled rovers loaded to the gills with Defenders pointing in every direction.

Just my two cents. -Manny

Whoa! At no point, will wheels/tracks be allowed. As technology becomes more advanced and affordable I can see the rules being tailored from year to year to make Quads less desirable and bipeds more desirable to builders. Trade-offs will also be important, and at some point when we have enough density to support it, I can see us having 2 weight classes (as RX-64 based bots would be signifigantly more powerful than an AX-12 or HS-5990 based bot, and we really don't want the competition to require $5k to compete).

What I would like to see though, and it will only happen over time and through collaboration, is our mechs (especially bipeds) become faster walkers. I'm not saying we ever want to see a walker cross the 15 foot arena in 2 seconds, but watching anything move at < 1in/sec is just painful. The reality is, the technology level is quite high to build a mech -- it involves quite a bit of robotics, electronics and software to make a walking robot that can be wirelessly controlled. There are two main technologies I see us really developing through this contest: the first is our telepresence systems including better wifi cam and control methodologies, and the second being the actual walking robot technology itself. Another technology that could come out of this is low-cost/lightweight ways for mechs to help steer themselves, etc....

-Fergs

mannyr7
06-22-2009, 03:00 PM
I see your point - but we do have to exist in the real world. And in the real world, I didn't see "a lot" of bipeds actually able to compete this year. And that's not a slight to the bip builders - it's a really tough problem!

Like it or not, it's the audience that ultimately pays the bills. If there are only a handful of people being entertained by the event come year two or three, I suspect RG isn't going to be too excited about sacrificing all that real estate. It has to draw a crowd. RG is about the door fees, and to collect the door fees, they have to deliver mass-appeal entertainment.

Personally, I appreciate that Dave and Simone were willing to take a flyer and permit us such a huge amount of floor space for a first-year, totally unproven, event. I think we need to give some thought to how we provide them maximum ROI on that.

Point taken. I'll get to work on my flamethrowers!:wink:

DresnerRobotics
06-22-2009, 03:18 PM
We're also going to be exploring the possibility of giving bipeds an HP advantage to balance them against quadrapods. Again, a large part of Mech Warfare is the engineering aspect, so I want to continue to encourage people to build bipeds even if they aren't as practical as a quad (because in many ways, they aren't). If that means that biped builders get a slight handicap as a bonus for building a more complicated mech, than so be it... its all about balance and making sure our competitors have fun.

Wheeled bots will never be an option as long as I'm in charge. At most, we might suppliment a battle with a LOW hp airsoft tank, just to add some variety... but that would be as a support role, not something someone was building to compete with.

darkback2
06-22-2009, 03:24 PM
I figure it will be really hard to strike a balance between indulging the audience, and indulging ourselves. Take for example bipeds vs quads. Obviously quads make this event easier to do than bipeds, and to be honest i'm not sure I would participate if it was for bipeds only (more because I believe in the efficiency of having bots that serve multiple purposes...which might mean a biped class, which in the end may result in very few participants. Lets face it, there were only 7 competitors signed up for the event, and really only 4 could participate. I know that had a lot to do with it being year one, but we had a year to make it a reality. Again, I think the fact that anything happened at all is incredible. And I really praise everyone who made an attempt at it.

I know this isn't exactly the right forum...IE the last 4 or 5 posts should probably be moved over to the other thread on post mech warfare analysis...moderators feel free to move this...

Scenario 1, Teams where each team is allowed to have a certain number of legs...so 4 bipeds vs 2 quads or a biped and a hexapod vs 2 quads and so on.

Scenario 2, Wheeled vehicles which are strictly defensive. Two teams, each with several mechs and a wheeled vehicle...the wheeled vehicle would represent the teams "captain" you score hitpoints by hitting the wheeled vehicle, hitting the mechs disables them for a few seconds...or hitting a mech takes it out of the competition, but does not actually score points for your team.

Scenario 3, One team has to defend the center position or some other position, and a larger force has to take it...the event would be based on the time it takes to establish control of the "castle"

I guess the point is, we could increase the amount of action simply by increasing the number of robots in the field...adding more cover, or varying cover might also help liven things up...as well. To be honest, people just like seeing the guns firing regardless of what was going on.


Seriously though...thanks to andrew, fergs, and david for puting on an amazing event.

OK...enough from me.

DB

Connor
06-22-2009, 03:42 PM
As someone who built a biped, only to have it not walk, I can say this much. It CAN be done. My Mech ALMOST walks, and with more time, I could have gotten it to work. Much of my problem was my choice of servo's, I used 5995's which are as strong as 5990's, but don't have heat sinks, and don't have thermal shutdown protection. I blew a total of 5 5995 servo's I believe in the course of building and competing with my mech. If I had of went with the tank guns vs the Defender guns, It probably would have worked alot better. I think the following needs to be done to make a scout based mech work.


1) Move the turret so that it's over the HIPS. Loose the forward part of the scout torso (or keep it for mounting light weight electronics, I used it to house the Fergy's target board

2) Must use 5990's for all the leg servo's except for hip rotation (turning).
3) Must have a MCU, using a Xbee(or bluetooth) + SSC32 alone, will not work in Robogames environment.
4) Maybe move the batteries to the feet.. I'm not 100% sure this is necessary, or will help. Would need to be tested.
5) Switch out the guns to the lighter weight tank guns.

FYI, My turret alone weighed in around 1lb or better. That wasn't counting the pan/tilt servos.. Just the camera, 2 guns and the ED 209 Chassis. My 5200mAh LiPo pack was around 10.2oz.

Thanks, Connor

Thanks, Connor

Adrenalynn
06-22-2009, 03:54 PM
>>using a Xbee(or bluetooth) + SSC32 alone, will not work in Robogames environment.

I disagree with you here. It's implementation. Landshark went all over robogames. Picked up blocks and handed 'em to children, untied people's shoe laces, even killed himself with a screwdriver. Sometimes >100ft - nary a problem.

I'd like to chat with you about your implementation because I never had even a single burp.

lnxfergy
06-22-2009, 04:02 PM
>>using a Xbee(or bluetooth) + SSC32 alone, will not work in Robogames environment.

I disagree with you here. It's implementation. Landshark went all over robogames. Picked up blocks and handed 'em to children, untied people's shoe laces, even killed himself with a screwdriver. Sometimes >100ft - nary a problem.

I'd like to chat with you about your implementation because I never had even a single burp.

How much data were you sending though? The auto-resend capability of paired XBEEs can handle quite a bit of interference... but if the amount of data gets too high, it will run out of time to get everything across safely... If I recall, Connor was sending entire poses through the XBEE? Not sure how often the arm controller updates the outputs of the 3(?) servos on Landshark though...

-Fergs

ScuD
06-22-2009, 04:22 PM
Obviously I didn't attend the games, nor have I ever before. But I do know that a crowd likes stuff getting destroyed.

I'm not saying mechs getting destroyed, but since some of the competitors seem to be pro more cover, why not replace the cardboard boxes with boxes made of uhm... I dunno what it's called, back here it's known as "architect's cardboard", sort of a foam substrate lined with paper. Easy to cut into squares, glue together with elmer's glue and some pins, and you've got a box that takes on more visible damage than a cardboard box.

Make a simple cardboard template with some rectangles cut out, spray paint black over gray and you've got a building.

My brother's into wargames &#224; la games workshop and builds these things by the dozens in a few hours, plus it's cheap.

come to think of it, might be better if this stuff got posted in the "Mech warfare year one recap" thread.. sorry bout that..

Adrenalynn
06-22-2009, 04:23 PM
Sorry - this is pretty OT...

Five, with motor control. That's part of the implementation. Send command, wait 20ms, send it again, wait 20, send it again, wait 20, rinse, lather, repeat.

Tip #2: Lower the baudrate, and nestle the commands into a single line. Don't go more bps than you absolutely need, and consider the packet structure carefully when deciding what that actually is. Don't underestimate the importance of this tip! Any time you have a dirty "wire" dropping the baudrate is going to actually get MORE data through.

2.1: Send multiple carriage returns. Some message arriving is generally better than feeding it nothing. There is either actionable data, or garbage.

Tip #3: Get it higher, get it less obstructed. I had serious space constraints given the rules, but the XBee was mounted on [microwave nearly-transparent] plastic above everything else with a clear look from 360deg. The arm side was of course elevated. Get the computer side off the table... Consider the ground absorption of microwave radiation.

10 servos, worst-case, is 71 bytes/all-move command.
9600bps = 1200byte/sec

1200byte/71 = 16.9 frames/sec.

So even at 9600 baud, you're exceeding the update speed of your video, typically. I'd be willing to bet he wasn't at 9600... If you don't do the math, 9600 seems excessively slow.

Connor
06-22-2009, 04:33 PM
Sorry - this is pretty OT...

Five, with motor control. That's part of the implementation. Send command, wait 20ms, send it again, wait 20, send it again, wait 20, rinse, lather, repeat.

Tip #2: Lower the baudrate, and nestle the commands into a single line. Don't go more bps than you absolutely need, and consider the packet structure carefully when deciding what that actually is. Don't underestimate the importance of this tip! Any time you have a dirty "wire" dropping the baudrate is going to actually get MORE data through.

2.1: Send multiple carriage returns. Some message arriving is generally better than feeding it nothing. There is either actionable data, or garbage.

Tip #3: Get it higher, get it less obstructed. I had serious space constraints given the rules, but the XBee was mounted on [microwave nearly-transparent] plastic above everything else with a clear look from 360deg. The arm side was of course elevated. Get the computer side off the table... Consider the ground absorption of microwave radiation.

10 servos, worst-case, is 71 bytes/all-move command.
9600bps = 1200byte/sec

1200byte/71 = 16.9 frames/sec.

So even at 9600 baud, you're exceeding the update speed of your video, typically. I'd be willing to bet he wasn't at 9600... If you don't do the math, 9600 seems excessively slow.

I was at 38400. I would send multple commands to the SSC-32 in one string (Move both pan/tilt servo at the same time). Problem wasn't always with the send, it was with querying the SSC-32 to figure out where the player was so I could issue a stop command at the correct place. Also, had issues with the dropping the end fire command, ended up having to double/tripple tap the fire button to get it to stop. Other issues was my pan/tilt would studer or, would turn real quick in one direction and then continue on the correct direction (dropping a number from the usec command or something) I think some modification to the SSC-32 code to include some Error checking/correction would go along way to making it more stable and usable with a Xbee. Maybe we make a SSC-32 Firmware version for Mech Warefare and standardize it for people or something. ??

Thanks, Connor

Adrenalynn
06-22-2009, 04:33 PM
Arduously moved from: http://forums.trossenrobotics.com/showthread.php?t=3352 pages #4 and #5 into http://forums.trossenrobotics.com/showthread.php?t=3335

Split at:http://forums.trossenrobotics.com/showpost.php?p=31653&postcount=34

Adrenalynn
06-22-2009, 04:37 PM
How many servos did you have going on? 38400 seems like 3x too much data with at least 3^2 the potential for loss.

Did you really need to echo every servo?

I looked at the firmware. It's written against an old version of the toolchain, won't compile against the modern toolchain, and porting would be a beach. I think I'd take a pass myself - I'm looking to get away from PWM - and if I weren't, I'd be building my own dedicated controller.

Connor
06-22-2009, 04:43 PM
How many servos did you have going on? 38400 seems like 3x too much data with at least 3^2 the potential for loss.

Did you really need to echo every servo?

I looked at the firmware. It's written against an old version of the toolchain, won't compile against the modern toolchain, and porting would be a beach. I think I'd take a pass myself - I'm looking to get away from PWM - and if I weren't, I'd be building my own dedicated controller.

Only 2 Servos (Pan/Tilt), Also Every 1/2 Second I would query the system for voltage (if no commands were in the queue), 2 Channels for guns (Start/Stop when Pulling trigger), and start/stop commands for the player. I didn't echo every servo, I would just query the player rather quickly (and continuously) when I would change from forward/reverse to stop, or when I would need to issue a turn while moving.

Thanks, Connor

Adrenalynn
06-22-2009, 05:51 PM
How many channels total were you running? 8?

Sienna
06-23-2009, 05:41 AM
Re: Audience Engagement
If audience engagement is that important, why not do small easy things, such as have a optical interupt sensor at the barrel's edge, that causes some amber or red-orange LEDs to light up near the barrel's edge to simulate the flash of firing a gun? And then tie the hit system into a sequence of red / orange / yellow LEDs, simulating a fire. This would provide indication to the audience that the gun is firing, even if they can't see the actual BB. (And done correctly, this would also provide feedback to the mech operator, as I think most cameras can see their own gun. When the barrel stops flashing, you know your outta ammo, or your gun jammed)

Re: Quads vs Bipeds
(Ignoring the fact that tripods are almost universally discounted :P)
I finally wised up to the fact that I am not really ready for a biped, especially for something like mechwarfare. So, this past weekend I stripped my two bots (my biped and tripod) down into a pile of parts each. I rebuilt that pile into a small quad for gait development, and have enough servos left over to build a second quad for competition. Normally, software projects cause me disinterest (which is odd if you consider my degree is in CS), so my bots usually get to the built mechanically stage and then collect dust. I am hoping I will find the motivation I need to actually do the software work on the quads, so I can compete next year. But that is all extraneous background info. What I am trying to say is that because of the sheer amount of multidiscipline work that is involved in getting *anything* walking, maybe competition handicaps should be based on "veterancy"? For instance, the first year you compete, you are allowed to bring in a quad, and you will not be penalized. If you bring a quad your second year to the competition, your hitpoints decrease by 25%. If you bring a quad in your third year to the competition, your hitpoints decrease by 50%. At any point in time, you may bring in a biped, which when competed would have the full hitpoint allocation. Some sort of "sliding scale" like this I think would encourage people to compete at the entry level, while allowing (encouraging) them to "grow" up to a biped.

Re: Video
I would hate to see some rule written to allow only IP camera systems. That is of course self serving, as I happen to have a very lightweight (but high power draw :() video transmission system in the 900MHz range. (And because the competition is over, and people are moving to trendnet cameras anyway, I will say I got it here (http://www.rangevideo.com/index.php?main_page=index&cPath=35_21&zenid=0611e8dee5983ed8fa79cb5a3611870e).) Anyway, I think there are viable solutions to wireless video that are non-IP. If you want to limit 2.4GHz systems to avoid interference with all the bluetooth, zigbee, wifi, etc, I guess I could see that, but unless people are moving to 900MHz zigbee as well, I would hate to be forced into a solution that requires me to have a full high power laptop setup at the competition.

lnxfergy
06-23-2009, 09:07 AM
Re: Audience Engagement
I like the barrel idea, quite a few of the mechs had an LED interfaced the scoring board, but it wasn't neccessarily visible -- plan to change that in future years. I'm not really in favor of adding additional 'requirements' to bot builders, but I think the barrel light would have enough 'cool' factor that most people would want one...

Re: Quads vs Bipeds <(Ignoring the fact that tripods are almost universally discounted :P)>
Hey, you're the tripod builder, and you just said you were dumping it in favor of a quad! The rules currently mention 'tripods', we just don't have anyone building them...

I don't think our goal is ever to eliminate Quads, but we have to craft the rules a little bit (especially right now) to encourage the development of Bipeds.. they have to at least be competitive (which right now, they really aren't). I'm not really a big fan of the "veterancy" setup though.... it may take some people several years just to get a reliable quad built, why would we force them to switch to a biped...

Re: Video <I would hate to see some rule written to allow only IP camera systems.>
Adrenalynn did some onsite testing with some fairly high-end non-IP cameras (see the 2009 Comptetitors Recap thread for the actual results). Her results were basically that non-IP cams just won't give you anything to look at... Even some of the IP cams failed to work onsite (see gdubb's comments on the AirLink).

-Fergs

darkback2
06-23-2009, 09:55 AM
Why do we have to encourage bipeds? Is it because they are harder? Or because they look more like what we think mechs should look like?

IMHO quads actually worked. Mech warfare is hard enough as it is. If you want, have a biped class where people who build bipeds can compete against each other in a fairer fight...But I say if quads work then quads work. Lets all remember bipeds for this event will come at a HUGE cost. Bipeds mean upgraded/more expensive servos, and lighter/fewer guns...(clyde being the exception) Bipeds already have a target plate advantage...(you can't shoot them from the sides = no strafing)...take away the penalty for falling down...maybe massage the hitpoints if you really want...

For me...a biped mech would be a 4 to 5 year project budget wise...If I were to drop all other projects, and I'm not ready to do that.

Oh...and sienna...remember, my first attempt at a mech was a tripod. of sorts...a sucky one...but a tripod none the less.

So...Not that anyone is taking a vote...but thats mine.

lnxfergy
06-23-2009, 10:14 AM
DB -

The reason to encourage bipeds is apparently because of the game (personally, I have no beef with quads.. that said, I'm planning to have a quad and a biped next year if I can swing it).

I'm not saying that bipeds are going to be the rule of the land... in fact, I personally don't see a biped taking a podium even for some time... technology dictates that we will have a bunch of quads, and a few slower bipeds. The reduced number of target plates and a small HP bonus is probably enough to at least level the field so that an awesome biped builder could podium...

I'm definately not in favor of removing the fall over rule, since quads are gonna be scurrying around more and thus running into walls more (which also sets off the plates... its sort of a quad version of falling over) -- plus, nobody wants to sit there and have to pause over and over again to pick up a biped that can't walk!

We just don't have the density yet to have separate leagues.. maybe some day.

-Fergs

DresnerRobotics
06-23-2009, 11:06 AM
Why do we have to encourage bipeds? Is it because they are harder? Or because they look more like what we think mechs should look like?

IMHO quads actually worked. Mech warfare is hard enough as it is. If you want, have a biped class where people who build bipeds can compete against each other in a fairer fight...But I say if quads work then quads work. Lets all remember bipeds for this event will come at a HUGE cost. Bipeds mean upgraded/more expensive servos, and lighter/fewer guns...(clyde being the exception) Bipeds already have a target plate advantage...(you can't shoot them from the sides = no strafing)...take away the penalty for falling down...maybe massage the hitpoints if you really want...

For me...a biped mech would be a 4 to 5 year project budget wise...If I were to drop all other projects, and I'm not ready to do that.

Oh...and sienna...remember, my first attempt at a mech was a tripod. of sorts...a sucky one...but a tripod none the less.

So...Not that anyone is taking a vote...but thats mine.

Why encourage bipeds? Because this is as much of an engineering competition as it is a combat competition. If you take the harder challenge, there should be at least some incentive to do so. We're looking to make this more inviting for people, including bipeds is a must, and if we need to make adjustments for balance, then we will.

Its really no different than a multiplayer videogame, if a certain 'class' is distinctly more powerful than another, we adjust the game rules to balance it accordingly. We want to challenge people, we want to have fun.


Also: Nobody said anything about forcing the use of Wifi cameras. Even the suggestion of a centrally managed wifi video system will be optional. Thing is, if we end up with a working video solution and you choose to pave your own road, later years will not be as forgiving if your video doesn't work.

dcalkins
07-07-2009, 02:52 PM
Does anyone have some large (3000x2000 pixel or higher) res action photos I can use?

Please do NOT email them to me. Just point to a website. Full credit will be given.

Adrenalynn
07-07-2009, 03:51 PM
I'll be in front of the computer in a couple hours and can pull something for you then, I'm sure. I was shooting 8.25mpixel, so even if we a little crop is needed, we can get your 6mpixels.

societyofrobots
07-11-2009, 06:44 PM
I think the bipeds not working thing was a symptom of procrastination, and not because bipeds maybe inherently suck. Quads do seem to have an advantage, though, based on last months competition. But if it wasn't for one particular quad running circles around everyone, we'd probably all say quads sucked too =P

If we get enough confirmed competitors next year, maybe there could be separate classes . . . like a biped class, or a weight-based class, etc. And if we do multiplayer games, perhaps a requirement for a quad/biped balance between teams. Just thoughts . . .

lnxfergy
07-11-2009, 09:59 PM
I think the bipeds not working thing was a symptom of procrastination, and not because bipeds maybe inherently suck. Quads do seem to have an advantage, though, based on last months competition. But if it wasn't for one particular quad running circles around everyone, we'd probably all say quads sucked too =P

If we get enough confirmed competitors next year, maybe there could be separate classes . . . like a biped class, or a weight-based class, etc. And if we do multiplayer games, perhaps a requirement for a quad/biped balance between teams. Just thoughts . . .

I think the bipeds not working has to do with:


How seriously difficult it is to balance a heavy biped
Real world budgets for hobbyist roboticists
Real world time constraints for hobbyist roboticists (who also hold jobs... note that I was only working 20hrs/wk during the 2 weeks I built Issy, and that was 20hrs/wk at a Uni during the summer.. yeah, most people work more than that).
That most people who built these were also building their "first walking bot"
Limited budgets.. again, cause yeah, I don't think any of our mech builders had MechEng degrees, meaning they designed mostly be experimentation (I know I did), and experiments are costly.

As for quads having an advantage... how can we say? We have yet to see a match with a walking biped versus a walking quad (Don't feel bad there Connor, I just slammed DB as much as you... :veryhappy:). Seriously though, building a biped, on a limited budget, that carries a crud load of mech gear, and runs for 15 minutes is serious challenge. I'm frankly amazed how far many of our builders made it given thier limited time/budget/experiance-with-walking-bots. We also have a full year this time around, whereas I don't think most people were building mechs before Jan/Feb (except Bheka, which seems to have to been finished way back in like October.. of 2004 :veryhappy:)

We will most certainly have multiplayer games next year (we're overhauling the scoring system to handle it, among other upgrades) -- but I doubt we'll have the density to support separate classes yet.

-Fergs

societyofrobots
07-11-2009, 10:47 PM
I guess being a mechanical engineer helps me keep costs really low =)
(but makes debugging code a huge pain, at least code is free! hehe)

Anyway, I'm expecting the same bipeds enter next year, but work this time and do impressive stuff =)

but I'm also expecting no more than ~6 working bots, tops . . .