Author Topic: Robot combat test (Unity engine)  (Read 3064 times)

Offline Tashic

Robot combat test (Unity engine)
« on: August 04, 2016, 03:42:58 PM »
I've experimented for about a week trying to make something resembling a robot using the physics engine and this is the result:https://gametechmods.com/uploads/files/9548Test.rar

The main thing that I don't like about the unity engine is that there isn't a cylinder collider, and instead I had to use a mesh, and because a mesh isn't completely smooth, I had to bodge a rubber like wheel, adding complexity, and it's still not very smooth ( without this it would be like using a cube as a wheel ).

And the wheel collider that Unity  provides is useless for anything other than cars.

If someone has used Unity and knows a better way of doing wheels for this application, let me know.

Offline R01

  • Heavyweight
  • Posts: 771
  • Rep: 1
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #1 on: August 04, 2016, 03:47:59 PM »
I haven't tested it nor do I have big unity knowledge, but wouldn't it be possible to add a smoothing group to the mesh, giving it better collision?

Edit:
Tested it, looks nice but it sadly has the same problem driving backwards like Robot Arena where it will drift into one direction. The top texture seems to be a bit blurry when far away(and I don't know why there's a hole in that top armor outside of the safety link).
Nice pick for the robot, did you create it or is there a 3d model floating around there?
Can we know the controls? I've tried pressing some buttons to see if I can use the weapon and I think one of them did something.
Other than the reverse movement and texture problem it does look nice.
Tournament History:
Showcases:
https://gametechmods.com/forums/index.php?topic=18882.0
https://gametechmods.com/forums/index.php?topic=19197.0

Offline Tashic

Re: Robot combat test (Unity engine)
« Reply #2 on: August 04, 2016, 03:54:29 PM »
Yes, physics will make going backwards difficult (same can happen in IRL if I'm not mistaken), but the unity physics feel a bit wonky in places.

Also controls are wasd/arrows for direction and f to turn on/off the spinner.

Model was made by me in sketchUp, then skinned in 3d studio max based on this photo:

Offline RedSawn

  • aka RedlineM203
  • *
  • Posts: 1939
  • Rep: 4
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #3 on: August 04, 2016, 04:28:20 PM »
Well it's certainly more convincing physics-wise then RA3. Can really sense the weight of the spinner impacts. It's like watching Warhead.




Offline Badger

  • Permanent Artifact
  • Giga Heavyweight
  • Posts: 6292
  • Rep: 3
  • Go ahead and bark after dark
  • Awards Donated money for site hosting 2019
    • View Profile
    • Awards
  • See profile for gamer tags: Yes
Re: Robot combat test (Unity engine)
« Reply #4 on: August 04, 2016, 04:30:11 PM »
This is so cool! Will you be working on it further?
also lol at most toxic guy around calling others out on this sh**

Offline Tashic

Re: Robot combat test (Unity engine)
« Reply #5 on: August 04, 2016, 04:41:17 PM »
I'm going to making other types of robots to test  ( rammers, flippers, VS.. ), then seeing if it's worth the effort, as it seams it will get pretty heavy in processing power.

Next robot I want to make is Original sin, this will test 4wd, hinged wedglets and pushing power.

Offline R01

  • Heavyweight
  • Posts: 771
  • Rep: 1
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #6 on: August 04, 2016, 04:54:37 PM »
Well it's certainly more convincing physics-wise then RA3. Can really sense the weight of the spinner impacts. It's like watching Warhead.
Well there goes my chance at a RA3 joke. Was going to say something like, new guy does a better game and physics engine in the same game creation software in a week than a professional company, but thought that it wouldn't really be needed/couldn't be fully compared since this has finished robots(much like the Robot Wars games) instead of custom built bots.

I'm going to making other types of robots to test  ( rammers, flippers, VS.. ), then seeing if it's worth the effort, as it seams it will get pretty heavy in processing power.

Next robot I want to make is Original sin, this will test 4wd, hinged wedglets and pushing power.
Yeah, game creation software isn't fully performance optimized but it should hopefully still work out.
Tournament History:
Showcases:
https://gametechmods.com/forums/index.php?topic=18882.0
https://gametechmods.com/forums/index.php?topic=19197.0

Offline Impossible343

  • Beetleweight
  • Posts: 136
  • Rep: 1
  • 5?
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #7 on: August 05, 2016, 03:59:20 AM »
Wow, this is amazing! I love how realistic the physics feel. Are there plans for any other RW bots?
Sometimes this game is on crack.
and now craaig knows how sakurai felt

Offline Philippa

  • The Queen of GTM
  • Giga Heavyweight
  • Posts: 6195
  • Rep: 0
  • ✊ Viva la Standard! ✊
  • Awards BOTM Winner
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #8 on: August 05, 2016, 11:55:31 AM »
Nice, is there a way to change the camera angle or is it fixed? Pretty cool looking Carbide too.

Offline Jules

  • *
  • Posts: 171
  • Rep: 2
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #9 on: August 06, 2016, 05:36:15 PM »
I dig it.

Do you happen to know what weights you have the chassis, the wheels, and the weapon set to?

Offline Tashic

Re: Robot combat test (Unity engine)
« Reply #10 on: August 06, 2016, 05:48:54 PM »
Right now they are: body and blade to mass 1, wheels 0.5.

I'm also wondering if the mass is the density of the object that with the volume is calculated the actual total mass, or if it is the total mass of the object regardless of volume.

Offline Jules

  • *
  • Posts: 171
  • Rep: 2
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #11 on: August 06, 2016, 06:36:45 PM »
Cool. I am 99% sure it is total mass.

For example a 10x10 cube with a mass of 1 would weigh the same as a tiny cube with a mass of 1

Offline Badnik96

  • tired of your shit
  • *
  • Posts: 17683
  • Rep: 2
    • Badnik96GTM
  • Awards BOTM Winner
    • View Profile
    • BattleBots Wiki
    • Awards
  • See profile for gamer tags: Yes
  • Skype: Badnik96
Re: Robot combat test (Unity engine)
« Reply #12 on: August 07, 2016, 01:50:31 AM »
This is fantastic already. If you need help modeling, I am quite proficient in Rhino 3D and could provide some robots for you. I did a pretty nice rendition of Tombstone a while back if you need a frame of reference.

Please continue work on this, at the very least it could be a modern version of BotApp to fool around in.

Offline Tashic

Re: Robot combat test (Unity engine)
« Reply #13 on: August 12, 2016, 01:06:40 PM »
News:

Did some more work on this, changed a fiew things, added a new robot, more cameras, and added some textures and lighting for the arena (to look more like a studio arena).

I think the next robot I'm making is a flipper.

Download:https://gametechmods.com/uploads/files/2038Robot test 2.rar

Offline R01

  • Heavyweight
  • Posts: 771
  • Rep: 1
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #14 on: August 12, 2016, 04:22:58 PM »
Have you changed the physics? Carbide flips itself over when hitting the sphere and bounces a bit around after that(even if the blade doesn't touch the ground), nice to see original sin and it's effective wedge design. Now I want to see Sewer Snake.
Tournament History:
Showcases:
https://gametechmods.com/forums/index.php?topic=18882.0
https://gametechmods.com/forums/index.php?topic=19197.0

Offline Tashic

Re: Robot combat test (Unity engine)
« Reply #15 on: August 12, 2016, 04:40:43 PM »
Carbide flipping itself hitting the sphere is because it gets forced downwards and the blade then hits the ground.

Also I made sure that both robot's weights add up to 10 units of mass (100 doesn't work well), and both the sphere and cube are also 10 mass.

Offline Philippa

  • The Queen of GTM
  • Giga Heavyweight
  • Posts: 6195
  • Rep: 0
  • ✊ Viva la Standard! ✊
  • Awards BOTM Winner
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #16 on: August 12, 2016, 05:23:32 PM »
A very crude version of Behemoth could be used to test a moving joint on top of another moving joint. (The axe on top of the scoop.) I don't know if that's something that Unity would handle smoothly or not, so I could be talking out my ass for all I know.

I'll have a look at the update tomorrow, but I imagine it'll probably be pretty on point like the first one.

Offline R01

  • Heavyweight
  • Posts: 771
  • Rep: 1
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #17 on: August 12, 2016, 06:08:10 PM »
Carbide flipping itself hitting the sphere is because it gets forced downwards and the blade then hits the ground.

Also I made sure that both robot's weights add up to 10 units of mass (100 doesn't work well), and both the sphere and cube are also 10 mass.
Why does 100 not work? Also guess the objects weren't changed for this?
Tournament History:
Showcases:
https://gametechmods.com/forums/index.php?topic=18882.0
https://gametechmods.com/forums/index.php?topic=19197.0

Offline Tashic

Re: Robot combat test (Unity engine)
« Reply #18 on: August 12, 2016, 06:33:50 PM »
100 didn't work because joints started to be too floppy, and I already increased the calculation speed, and I don't want to increase it too much.

Offline RedSawn

  • aka RedlineM203
  • *
  • Posts: 1939
  • Rep: 4
    • View Profile
    • Awards
Re: Robot combat test (Unity engine)
« Reply #19 on: August 13, 2016, 02:23:37 AM »
Yeah - whatever info I got on Unity physics says to some effect joints where the weight difference is a multiple of 10 or 100 aren't going to function particularly well. Which is great for robot combat simulation when an AW foam wheel makes up 2 grams of a 150g UK ant.