Autonomous Combat League

[PressureLine]

Just found this in my blueprints:





Made by tob.s it's a second version of his 'erazer' bot. Incredibly fast, armed with three lasers on the front and a ramming attack that has to be seen to be believed. No idea what the partcount is, but it's gotta be around 100.

[MarvinMan]

I don't remember that one. Could you put the blueprint up? If its charging attack is as bad as you say, it should make for a decent opponent to test a collision avoidance system against.

[PressureLine]

Well it has a ground speed of 40-50, so it's ramming/charging attack usually results in the target being thrown a good distance (if the lasers don't cut it in half) when it hits. I'm not sure how complete the code is, since it doesn't seem to track the target very well, mostly concentrating on hooning around the arena looking awesome.

[MarvinMan]

Here's a sample of what I've been working on- the RedEye. The mechanical structure and navigation systems are done, and the weapons system makes an attempt at hitting its target. In my testing, it managed to do a reasonable amount of damage to the PressureTank, but was usually disarmed or disabled before immobilizing it.

The turret still needs work as it only covers a 180 degree range (it least it's the front 180 degrees this time), doesn't make any attempt to lead its target and has no elevation control. It would probably be a good idea to add a collision avoidance system too, so it doesn't charge head first into the first laser bot it comes across.

I haven't been able to get a stable or accurate part count out of a camera, but I counted it at 51, which puts it in the unfortunate position of being the smallest and least armored bot in the heavyweight class.

Blueprint: https://www.dropbox.com/s/h4nivwg9rfypqdz/bp_redeye?dl=0

[PressureLine]

Looks good, can't wait to try it out. I will also update the directory post once I get home

[PressureLine]

I haven't been able to get a stable or accurate part count out of a camera, but I counted it at 51, which puts it in the unfortunate position of being the smallest and least armored bot in the heavyweight class.

FWIW I counted it at 51 too. It can join your original autoX3 (52) and Sawtooth (51) in the light-heavyweight category.

[PressureLine]

Almost finished a mediumweight ground bot. I just want to get a workable 'approach but don't ram' system working

[MarvinMan]

As much as the target following behavior makes bots engage more quickly and keep attacking, the ramming behavior does somewhat ruin it as the winner will usually be the bot that ends up on the bottom (or has lasers). At least with random walk behavior, you tended to get multiple encounters between multiple bots with relatively few collisions, even if it took a lang time to get a complete kill.

I'm planning to add a system to keep an optimal distance to the target, based on what the broadside had,  but I can see the situation getting messy for bots designed to go forward only, especially in tighter parts of the arena. I'm thinking that adding a second set of rear-facing hooks to do collision avoidance might let the bot get away from its enemy more easily. Kinda makes me wish I'd kept the whole reversible drive and navigation system.

[PressureLine]

New middleweight competitor!

HailStorm

Weighing in at 44 parts, HailStorm is a middleweight Plasma Shooter capable of tracking and engaging both ground based and aerial opponents.

Get HailStorm here.

[PressureLine]

As much as the target following behavior makes bots engage more quickly and keep attacking, the ramming behavior does somewhat ruin it as the winner will usually be the bot that ends up on the bottom (or has lasers). At least with random walk behavior, you tended to get multiple encounters between multiple bots with relatively few collisions, even if it took a lang time to get a complete kill.

I'm planning to add a system to keep an optimal distance to the target, based on what the broadside had,  but I can see the situation getting messy for bots designed to go forward only, especially in tighter parts of the arena. I'm thinking that adding a second set of rear-facing hooks to do collision avoidance might let the bot get away from its enemy more easily. Kinda makes me wish I'd kept the whole reversible drive and navigation system.

The system in HailStorm won't work for all bots, but works pretty well with it's turret design, I'm happy with it anyway.

[MarvinMan]

The HailStorm is a very close match for my RedEye, and the winner seems to depend on how they encounter each other. If the HailStorm gets line of sight from a distance or there is a significant altitude difference, it usually has the advantage. If the bots meet at closer range, the RedEyes very basic aiming is much less of a disadvantage and the Hailstorm tends to drive over it or roll, exposing its vulnerable underside.

One other thing I've noticed is that the turret assembly appears to be fairly easy to dislodge, although that could be because the RedEye can only really hit things high enough to line up with its own turret.

[PressureLine]

Did a quick 3-way between Erazer, HailStorm and RedEye:

Spoiler: show

And the final results:

RedEye: show

Erazer: show

And the winner:

HailStorm: show

I think RedEye got caught by Erazer's lasers early on which cost RedEye it's two front wheels. Hailstorm added a bit more damage, before Erazer returned to finish the job. after that was a bit of back-and forth cannon fire between Erazer and HailStorm until all of Erazer's weapons were destroyed.

[PressureLine]

So I added cannon elevation control to RedEye (was easy once I figured out the 'part' output flow) and this is the result:



(Yeah, thats HailStorm in tiny pieces, a turretless Canary and an undamaged RedEye)

[MarvinMan]

Wow, I wasn't expecting a properly set up turret to make that much difference. Without it, I was finding it being 50-50 whether the RedEye or Hailstorm would win, and that was in a relatively tight arena where the Hailstorm seemed somewhat prone to getting stuck or rolling.

Regarding the first round with the erazer, I had a similar result against the SawTooth. The RedEye normally seems to take losing an axle quite well, it's the loss of the motors and camera directly behind it that disable it as they deal with moving the turret and deciding whether to enable the cannons.

[PressureLine]

To be fair, Canary and HailStorm were already engaging each other when RedEye joined the party, and it's hard to keep track of which bot actually did the most damage when it's a fairly short-ranged gunfight, but RedEye definately fired the killing blow on at least one of the two (although I'm pretty sure it was on both)

I have been going through my old blueprints, and it's actually really cool to see how (at least for my bots) the code has evolved over time as I have built them.

Still trying to think of the next 'gimmick' for my next bot, might try and give grenades a go.