This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
Messages - kix
Pages: 1 ... 15 16 17 18 19 20 21 [22] 23 24 25 26 27 28 29 ... 168
421
« 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,
422
« 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%
423
« on: January 12, 2020, 06:55:17 AM »
New build coming today. I finally isolated and fixed the source of the problem with the gyro dancing and unrealistic behaviors when everything is broken off an axle.
TL:DR - The mass of the remaining axle will be set to 0.00001 kg in the simulation. This should prevent weirdness.
The full story:
[Bug Fix] Fixed a bug where the remaining mass of the axle alone was set to several kilograms when the last part attached to the axle is broken off. This was causing robots to gyrodance and jitter unrealistically. I also identified the source of the extra mass -- see "POTENTIAL BUG SOURCE" below.
POTENTIAL BUG SOURCE - I just noticed that the "Axle" subassemblies on motors and actuators comes with its own rigidbody. For the Ampflow A40-300 motors, this "Axle" rigidbody has a mass of 4 kg. I tried reducing the axle mass to something more realistic, like 1 kg. This was a TERRIBLE change. It made robots drive poorly, with the wheels bouncing and not gripping. Apparently we had set this default axle mass value years ago in the very early stages of the game and completely forgotten about it.
From now on, every time we create a motor with an axle, we need to make sure that the axle has sufficient mass, maybe 4 kg for a 110 kg robot? When everything breaks off except for the axle the mass of the axle will be set to 0.00001 kg.
Ayy the issue that has been haunting us for months has finally been resolved
424
« on: January 11, 2020, 06:04:55 PM »
So basically Imma just let everyone enter their bots and no longer have a cap on how many can enter... for now. I'll most likely be making cuts after the 31st. So please, everybody (Robo, Sploodle, Wambo, HardBot, etc), ENTER!!!
Does this mean there's no more limit on how many bots per person? Asking just to be certain.
Yes
425
« on: January 11, 2020, 06:00:30 PM »
Here's Rainbow Circus 2. Cyar, this is one of your creations, correct?
[ Quoting of attachment images from other messages is not allowed ]
No it was made by Minthemad
426
« on: January 11, 2020, 03:25:22 PM »
I have some ideas on how to do in-game grouping, but it is going to take a lot of work and troubleshooting to get things like collisionshapes, component breakage, hit points, center of mass, etc working correctly.
Honestly: collision shapes - Atm maybe not really necessary untill much later builds omponent breakage - If you would group parts, i guess you would have combined HP and once it reaches 0, the whole struture falls off hit points - Huh, thats the one thing that my knowledge is non existant. center of mass - i guess if you did that before grouping, that wouldnt be an issue
427
« on: January 11, 2020, 02:19:40 PM »
Problem with the tool is that personally i wouldn't allow custom components for my tournament. Unless there is a way to block it. Someone could enter a bit with low weight and extremely high hp count
I was thinking that you would set a material and the tool would automatically compute the weight and hp upon creation.
Idk, setting a material not ingame is kinda eh. There can always be a chance that the modded part is better than the ingame part. Again, this modding tool is great for motors and parts like that but for components i dont find it too good
428
« on: January 11, 2020, 02:03:07 PM »
I did a little performance testing for 3 robots using the Unity profiler. The times shown here are only significant insofar as they are relative to each other.
The complexity of the robot is almost directly proportional to the time spent in physics processing.
Rainbow Circus 2 is a shell spinner with 125 components. It is pretty much unplayable on my laptop. kix saw significant slowdowns on this robot on his computer, particularly as it passed through the flame pit and all of the colliders triggered collisions with the flame trigger volume.
Mental Breakdown is a simple vertical spinner with 32 components. It runs well on my laptop at 100 ticks/second. It slows down, but is still playable at 300 ticks/second.
HUGE Spinny is a "Tombclone" with 18 components. It also runs well on my laptop, slightly better than Mental Breakdown.
Rainbow Circus 2 has about 7x the part count, and spends about 10x the amount of time spent in the CPU compared to HUGE Spinny.
In order to make cool-looking robots we need to figure out a more efficient way to create visual complexity than simply adding components. The simplest way to do this is to get the RR2 Component Modding Tool (CMT???) up and running. This will allow you to create something really cool-looking in Blender, then pull it into Unity and have Unity create a single mesh collider for it. With such a tool in place we could reduce the number of colliders in Rainbow Circus 2's shell from 109 down to 1. I expect that this would reduce physics frame time by 90%.
(Image removed from quote.)
Problem with the tool is that personally i wouldn't allow custom components for my tournament. Unless there is a way to block it. Someone could enter a bit with low weight and extremely high hp count
429
« on: January 11, 2020, 11:17:37 AM »
Maybe also add a group function for currently built bots that use ingame components, because cmt is gonna mean that bots should be rebuilt
430
« on: January 11, 2020, 11:12:23 AM »
(Image removed from quote.)
EEEEEEEEEEEEEEEE
Yeah, I know. That's where the motor was when I had transparent chassis on. I couldn't tell it was sticking out. Download it and the damage it can do will blow your mind!
I dont tend to download bots that i dont like
431
« on: January 11, 2020, 09:22:45 AM »
Yup. You guys are right. There was a bug in the breakage code.
When all of the parts connected to a spinner (a.k.a. a single wheel) are broken off, it was supposed to set the remaining mass (a.k.a. just the axle) to have a mass of something just higher than zero. Instead it was setting the remaining mass to be equal to the original mass. This meant that the robot would shake and wobble like crazy when everything was broken off other than the axle itself.
New build coming soon...
Nice, how about the 500 tick lagging like crazy on certain bots, even tho the cpu usage was not full?
I watched the video, and it looks like it was only when you were in the fire pit.
It aint the flame pit. Its the bot im certain
432
« on: January 11, 2020, 08:43:00 AM »
Yup. You guys are right. There was a bug in the breakage code.
When all of the parts connected to a spinner (a.k.a. a single wheel) are broken off, it was supposed to set the remaining mass (a.k.a. just the axle) to have a mass of something just higher than zero. Instead it was setting the remaining mass to be equal to the original mass. This meant that the robot would shake and wobble like crazy when everything was broken off other than the axle itself.
New build coming soon...
Nice, how about the 500 tick lagging like crazy on certailn bots, even tho the cpu usage was not full?
433
« on: January 10, 2020, 05:58:52 PM »
Tested the build with 500tick and boy is it amazing. However
You can see some issues in the vid:
Its still a massive improvement
EDIT: What do the other 2 sliders do on the right?
PS. The smooth oscillation is actually just beta having lwered wheels for previous builds
434
« on: January 10, 2020, 04:02:30 PM »
New bleeding edge build is coming out tomorrow. I have added sliders to adjust the following:
Physics Tick Rate (100 - 500 ticks per second)
Physics Default Solver Velocity Iterations
Physics Default Iterations
I might try adding a few more sliders to see if we can figure out a fix to the jittering problem. I have noticed that if I increase the physics tick rate to 300 I the jittering in Beta seems to become a smooth oscillation back and forth. It just about doubles the time spent in physics per frame though. My baby laptop CPU is getting HOT.
Thats weird. The fix might not be that? What if the wheels have too much grip or something which cause it to jump? 500? OOOH i also want to see that
435
« on: January 10, 2020, 03:54:36 PM »
 EEEEEEEEEEEEEEEE
436
« on: January 10, 2020, 02:39:52 PM »
I noticed with the new steel floor there is this weird clanking that seems to be coming from that rolling pin. Here's a video, play it with sound:
[ Quoting of attachment images from other messages is not allowed ]
I second this
437
« on: January 10, 2020, 01:57:34 PM »
Tested the new build and beta does not seem to be doing better. I really dont know what is causing the jitteriness. Try looking at 2.01 build from April or so. The driving was spot on there
438
« on: January 10, 2020, 08:18:36 AM »
I'm getting increasingly suspicious that Mac and PC simulate the game completely differently.
The problem I have is that even without turning on the weapons of Roller or un-fixed Panic Attack, they just can't steer well enough to so much as manoeuvre around the arena.
It is possible, though I suspect that it might be more related to CPU/GPU load than PC vs Mac.
On my computer if I keep Roller's weapon off I can drive around slowly. It skids around more than I would like, but it is okay. It gets all bumpy within a second after I turn on the weapon.
Both versions of Panic Attack that you sent me drive perfectly fine. A little slow, but they are big robots.
The version of Mental Breakdown that kix sent me a while ago drives like an absolutely maniac. Super-maneuverable. It overshoots like crazy when the computer AI is driving.
This is what Roller and Panic Attack look like on my end.
Smooth simulation quality, but both bots are far more broken than you've described on your end.
But, one thing to note is that the test area doesn't break Panic Attack. Only actually going into battle does.
Try driving on the pit floor, it might work there
439
« on: January 10, 2020, 07:25:15 AM »
see some minor improvements, but noting really that stopped jittering completely
It does seem like the wheels have no traction and they jitter because of that. Now gearing down could help, but that also means that its gonna be way slower and more terrible to drive.
I will say it again. The pit lit part of the arena is the only part where you can drive the bots on, and they wouldnt jitter
You are right! The arena floor was set to generic "Metal", while the pit is "Steel". The only difference between the two materials is that "Metal" was set to use the minimum friction coefficient of the material, while "Steel" used the average.
I'm not seeing a difference in the jittering (which doesn't appear too bad on my computer), but am seeing a noticeable difference in traction. Beta definitely has more traction with the Steel floor. Maybe we need a "friction" slider in the physics tab? 
I should have a bleeding edge build up with the change in the next 10 hours...
As long as it stops the jittering, im more than happy. There are fair amount of people wit bots that jitter.worst case scenario, set the floor material to steel.
440
« on: January 10, 2020, 04:22:15 AM »
kix,
I just posted a new 09January bleeding edge build. I think this one fixes the jitter for Beta. All I did was changed the Default Velocity Solver Iterations in Physics Settings from 4 to 10. Let me know what you think!
http://www.robot-rumble.com/bleedingedgebuilds/
Well i do see some minor improvements, but noting really that stopped jittering completely It does seem like the wheels have no traction and they jitter because of that. Now gearing down could help, but that also means that its gonna be way slower and more terrible to drive. I will say it again. The pit lit part of the arena is the only part where you can drive the bots on, and they wouldnt jitter
Pages: 1 ... 15 16 17 18 19 20 21 [22] 23 24 25 26 27 28 29 ... 168
|
Show Posts