Poll

What do you like doing most in RA2/RA3/robot combat games?

Building robots
Local single player battles with manually-controlled robots
Local multiplayer battles  with manually-controlled robots (PvP with controllers or a shared keyboard)
AI-only tournament battles
Other?  Please comment below.

Author Topic: Robot Rumble 2.0 - Robot Combat Simulator - Under Development  (Read 187334 times)

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1460 on: January 12, 2020, 08:56:49 AM »
The problem with the gyrodancing didn't go away. I'm thinking that if the motors stopped after a wheel is torn off, then the glitch will be fixed.
EDIT: The problem is fixed, apparently I mistook the gyro created by the wheels for the glitch.

Awesome!  There is still pretty significant gyro caused by the remaining wheel due to the extra 4 kg mass added by the remaining wheel's axle.

I'm thinking maybe we need another slider to play around with the ratio of wheel mass / chassis mass.  Currently, with the extra 4 kg added to the axle, wheel mass is between 4-5% of chassis mass.

My hypothesis is that as you raise the percentage of wheel mass to chassis mass:

Pros: Joint stiffness would increase, reducing the annoying bounciness of wheels.
Cons:  Wheel gyroscopic effects would also increase, throwing the robot out of balance and potentially causing it to gyro dance on its side once the wheel gets going fast enough.

Only one way to find out.  Its slider time!!! :)

Offline powerrave

  • Giga Heavyweight
  • Posts: 5372
  • Rep: 2
    • View Profile
    • Awards
  • See profile for gamer tags: Yes
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1461 on: January 12, 2020, 09:04:49 AM »
I tried the circus bot again, but this time multiple in the arena at once.
So, 4 was the engine wanting to commit suicide. 3 went well until collision; then the frames drop. 2 was fine. Frame drops only happens when components went clipping into each other.
This is run on the bleeding edge released today.

Unrelated but something that has also been a thing. AI on own robots being reset to the default of the game after installing a new version of the game (drag & drop over the previous version). This is a bit annoying since some of my bots just no long use their hammer or flipper until I give them their AI again.
"Always be yourself, unless you're a loser"


Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1462 on: January 12, 2020, 09:30:15 AM »
I do notice my fans are spinning up a lot and then keeping that constant. Even in workshop or something.
Well from my experience, the fps is uncapped, and at uncapped fps, gpu runs at 100%

Offline powerrave

  • Giga Heavyweight
  • Posts: 5372
  • Rep: 2
    • View Profile
    • Awards
  • See profile for gamer tags: Yes
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1463 on: January 12, 2020, 10:24:13 AM »
When I checked performance monitoring, GPU ran at about 30 to 35%...
"Always be yourself, unless you're a loser"


Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1464 on: January 12, 2020, 11:23:20 AM »
I tried the circus bot again, but this time multiple in the arena at once.
So, 4 was the engine wanting to commit suicide. 3 went well until collision; then the frames drop. 2 was fine. Frame drops only happens when components went clipping into each other.
This is run on the bleeding edge released today.

Excellent.  This helps bracket the problem down.  I would like to get the game to the point where it defaults to 300 physics ticks per second and runs at 60 fps on a  midrange laptop with 4 AI driven opponents with 50 colliders each.  It is going to take a lot of optimizing to get there, but I think it is possible.

Unrelated but something that has also been a thing. AI on own robots being reset to the default of the game after installing a new version of the game (drag & drop over the previous version). This is a bit annoying since some of my bots just no long use their hammer or flipper until I give them their AI again.

Ah yes.  The AI rewrite.  It will happen, I promise!  Right now the AI system is just as bad as physics.  It takes just as much CPU time.  That is my fault, and I intend to fix it.  The fix will include reimagining how AI is saved and loaded.

Offline powerrave

  • Giga Heavyweight
  • Posts: 5372
  • Rep: 2
    • View Profile
    • Awards
  • See profile for gamer tags: Yes
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1465 on: January 12, 2020, 12:08:18 PM »
Good to know that this AI rewrite issue will be tackled eventually.
Should note that all the results with the circus bots was on 500 ticks.
"Always be yourself, unless you're a loser"


Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1466 on: January 12, 2020, 01:37:30 PM »
The problem with the gyrodancing didn't go away. I'm thinking that if the motors stopped after a wheel is torn off, then the glitch will be fixed.
EDIT: The problem is fixed, apparently I mistook the gyro created by the wheels for the glitch.

Awesome!  There is still pretty significant gyro caused by the remaining wheel due to the extra 4 kg mass added by the remaining wheel's axle.

I'm thinking maybe we need another slider to play around with the ratio of wheel mass / chassis mass.  Currently, with the extra 4 kg added to the axle, wheel mass is between 4-5% of chassis mass.

My hypothesis is that as you raise the percentage of wheel mass to chassis mass:

Pros: Joint stiffness would increase, reducing the annoying bounciness of wheels.
Cons:  Wheel gyroscopic effects would also increase, throwing the robot out of balance and potentially causing it to gyro dance on its side once the wheel gets going fast enough.

Only one way to find out.  Its slider time!!! :)
How about if a motor has nothing on it, it auto turns off,

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1467 on: January 12, 2020, 01:42:28 PM »
How about if a motor has nothing on it, it auto turns off,
I've been trying to say this for weeks!

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1468 on: January 12, 2020, 02:36:51 PM »
But then you don’t get the cool spinning axle with nothing on it!  :smile:

You might be right.  Lets see how this feels.  If we don’t like it we can always change it.

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1469 on: January 12, 2020, 02:37:59 PM »
I think the cool spinning axle is causing the glitching.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1470 on: January 12, 2020, 02:40:29 PM »
But then you don’t get the cool spinning axle with nothing on it!  :smile:

You might be right.  Lets see how this feels.  If we don’t like it we can always change it.
Tbh i dont think that people will notice cool spinning axle in a match

EDIT: Can we get an sparks slider? Atm I kinda miss the sparks on low impact.

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1471 on: January 12, 2020, 02:43:05 PM »
Yeah, you can’t even tell that the axle is spinning.

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1472 on: January 12, 2020, 03:03:02 PM »
I think the cool spinning axle is causing the glitching.

Sort of.  The root cause was the fact that the axle masses were outrageously high (4kg, making each wheel about 10% of the mass of the chassis) and spinning.  If the axle mass wasn't so high then when the wheel broke off the resulting mass would have been so small that the spinning axle wouldn't have created gyro dancing issues.

I think we might need the axle masses to be unrealistically large, but I can't be sure until we have had a chance to test different values on lots of different robots to see if it affects drivability.  Unity recommends that the mass ratio between two jointed rigidbodies not exceed 10:1.  If the wheel/axle combination is smaller than what Unity suggests that the resulting simulation might get unstable.  I put in a new physics slider on the next build to test this out.

For now, lets test to see if we need the extra axle mass.  Set the slider all the way to 1% and see if robots still drive okay.  If so then we can set the axle masses to more realistic values (around 1 kg for an Ampflow A40-300).

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1473 on: January 12, 2020, 03:04:37 PM »
EDIT: Can we get an sparks slider? Atm I kinda miss the sparks on low impact.

Yeah baby! :)

Sparks are so satisfying!  I know they aren't always realistic, but stupid and over the top are what we live for in robot combat!

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1474 on: January 12, 2020, 03:08:55 PM »
EDIT: Can we get an sparks slider? Atm I kinda miss the sparks on low impact.

Yeah baby! :)

Sparks are so satisfying!  I know they aren't always realistic, but stupid and over the top are what we live for in robot combat!
Yay! My sawbot doesn’t spark like it should, so I am all for SPARK SLIDER!!!!!!!!

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1475 on: January 12, 2020, 03:40:12 PM »
The spark slider didn't make it into today's Alpha build (coming soon!), but please remind me when I get back from my game dev hiatus next week!

The 12January2020 Alpha Build is up on itch.io:


https://robot-rumble.itch.io/builds


Here's what I have for the Alpha Build.  Basically it incorporates all the changes in the past few day's worth of Bleeding Edge Builds.  I also fixed the banging arena auger:

[Updates in the January 12th Build]

[Changed] Changed the arena floor material to "Steel".   This increases the traction of most robots.

[Added] Added sliders for physics tick rate. This can be used to tweak physics smoothness versus game frame rate.​  High end machines can support smooth play at 500 physics ticks per second.  Low end machines might need to reduce this to 100 physics ticks per second.

[Added] Added a Physics Mass Wheel Ratio Slider.  Default Wheel Mass Ratio is now set at 10%.​​  This is going to be using for testing driving behavior on a wide variety of robots.​

[Bug Fix] Fixed bug where spinner mass was still very large when all parts were broken off an axle. This was causing gyro dancing and uncontrolled behavior for many robots.  Now the mass is reduced to a very small amount. 

[Bug Fix] Raised height of Arena Auger so it didn't continuously bash itself into the floor, causing a continuous loud "bang" "bang" "bang" sound.​

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1476 on: January 12, 2020, 03:46:21 PM »
The spark slider didn't make it into today's Alpha build (coming soon!), but please remind me when I get back from my game dev hiatus next week!
Dang it... I really wanted to make an arc welder XD.

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1477 on: January 12, 2020, 03:53:56 PM »
And about breakable armor plates, can’t you just treat the plates as components?

Offline min440303

  • Lightweight
  • Posts: 192
  • Rep: 2
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1478 on: January 12, 2020, 07:49:49 PM »
The discussion has gone so fast.
For me the most satisfying spinner slider settings is 150%/150%/50%
Mad but not mad.

Stuff I kinda done at

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1479 on: January 13, 2020, 01:54:39 PM »
And about breakable armor plates, can’t you just treat the plates as components?

The chassis is special because it is the only component with separate armor plates.  It will need its own special set of rules for breakage.  Next week! :)