Welcome, Guest. Please login or register.

Login with username, password and session length

 
Advanced search

1411273 Posts in 69323 Topics- by 58380 Members - Latest Member: bob1029

March 28, 2024, 02:13:24 AM

Need hosting? Check out Digital Ocean
(more details in this thread)
TIGSource ForumsCommunityDevLogsWobbledogs
Pages: 1 2 3 [4] 5 6 ... 37
Print
Author Topic: Wobbledogs  (Read 229686 times)
CraigSnedeker
Level 0
**


View Profile
« Reply #60 on: March 16, 2016, 07:57:43 AM »

I'm really excited to play this, it looks so fun haha.
Logged
jctwood
Level 10
*****



View Profile WWW
« Reply #61 on: March 16, 2016, 08:29:23 AM »

Great explanation! Have been intrigued as to how you approached it from the beginning. You may have mentioned it already but I'd love to know if you are using vanilla unity physics and if so is it all hinge constraints? Voxel dog was a real surprise but I kind of love it!
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #62 on: March 16, 2016, 03:18:40 PM »

You should add a playable character that would be able to:
push the "pets"
attempt to ride the "pets"
and more i just dont have that much ideas right now.
(love the gifs though)

Glad you're enjoying the gifs! That's a fun idea, but I'm planning on more of a "god" style game, where you play a disembodied hand or something. A playable character would probably about double my scope and could make a lot of the interactions I'm planning kinda frustrating. Something to keep in mind though!

I'm really excited to play this, it looks so fun haha.

Awesome to hear!

Great explanation! Have been intrigued as to how you approached it from the beginning. You may have mentioned it already but I'd love to know if you are using vanilla unity physics and if so is it all hinge constraints? Voxel dog was a real surprise but I kind of love it!

Thanks! I don't think I've mentioned that anywhere, actually. I'm absolutely just using vanilla unity physics (though I've obviously tweaked a few of the global settings). I've run into a few issues with the system (namely that it's non-deterministic even during a single simulation), but it's still way better than anything I could make on my own, and it interfaces nicely with a bunch of built-in unity stuff that I'm relying on. I'm using a HingeJoint to connect the two body segments because that's such a simple thing, but all the legs are using ConfigurableJoints for their connections. They started as hinges, then I made them CharacterJoints, but eventually I moved to configurables because they surface all the properties of the other joints and allow me to use joint motors, which I ended up using as a sort of joint tension model.

---

Early update today but I gotta jet out soon. Mostly I've been designing a model for dog emotions and memory. I've just started the implementation and it's going ok, but it'll be hard to say for sure until I can run some tests.

The high level idea is that dogs will have a constant emotional state, and whenever a behavior finishes, the dog's brain will associate its current emotional state with that behavior type (I'll probably need to make this associate a change in emotional state from start to end of the behavior, but I'm starting like this for now). When deciding what behaviors to run, that information will be used along with the base emotions/utilties the behavior advertises under normal circumstances.

The current plan is that a dog's state will consist of:
  • Curiosity
  • Excitement/Fear/Anger
  • Hunger
  • Arousal

Curiosity will go from 0 to 1 as a dog reaches "child" age, and then will slowly decay back to 0 as it gets older.

Excitement/Fear/Anger is all a single statistic that switches context depending on curiosity. A super curious dog will treat that emotion like excitement, a only marginally curious dog will treat it as fear, and a non-curious dog will treat it as anger. I'm also allowing the three emotions to bleed together, so you can get a dog that's kind of excited and kind of scared.

Hunger is gonna be a simple value that decays over time.

Arousal is the one I might end up taking out. I'd like to get the dogs to mate with each other autonomously, but if it becomes too much work I could just make breeding a mostly UI thing. For now I'm moving forward with the idea of autonomous mating. This means I'm gonna have the arousal stat flare up semi-randomly during dog puberty, level out and contextually flare up during adult years, and slowly taper off entirely. Putting this stat into the system should hopefully give me a simple model for fetishes. A dog that consistently gets arousal flareups when eating might associate food with sex.

I'm really trying to make the model as simple as possible while opening up lots of possibilities. This strategy feels good to me so far, but we'll see!

I also spent a good deal of time tweaking mid-air rotations some more. It's fun to do and I want it to feel good. I just added a rule that says dogs can't correct against angular velocity unless their non-angular y velocity is low enough. This creates a nice little peak when thrown in the air where the dogs can try to level out, but if they screw it up before they start to fall back to earth, they just have to go with the flow of momentum.
Logged

Fuzzy Branch
Level 0
***



View Profile WWW
« Reply #63 on: March 16, 2016, 05:55:26 PM »

Cool project. I've been working on a similar kind of thing for a little while now, but with a bit of a different focus -- being able to create any kind of body-shape and teach them to do certain tasks by training their neural-net 'brains'. I've been scouring the internet for similar physics-based genetic-algorithm creatures for the past year or so and yours is the first one I've come across, so i'm very interested to see how your project progresses, especially with regards to higher-level behaviors like mating, play, anger, etc. since I won't be getting to that kind of stuff for a while yet -- focusing more on basics like locomotion at the moment.

I don't want to hijack your thread so I won't post info about my project here, but if you'd be interested in sharing information or lessons learned/observations let me know and I can send some links, maybe we could learn something from our different approaches.

Anyway, good luck with your pet dogs! Looking forward to seeing them progress.
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #64 on: March 17, 2016, 08:44:15 AM »

Cool project. I've been working on a similar kind of thing for a little while now, but with a bit of a different focus -- being able to create any kind of body-shape and teach them to do certain tasks by training their neural-net 'brains'. I've been scouring the internet for similar physics-based genetic-algorithm creatures for the past year or so and yours is the first one I've come across, so i'm very interested to see how your project progresses, especially with regards to higher-level behaviors like mating, play, anger, etc. since I won't be getting to that kind of stuff for a while yet -- focusing more on basics like locomotion at the moment.

I don't want to hijack your thread so I won't post info about my project here, but if you'd be interested in sharing information or lessons learned/observations let me know and I can send some links, maybe we could learn something from our different approaches.

Anyway, good luck with your pet dogs! Looking forward to seeing them progress.

Awesome! Yeah I'd love to hear more about it. Feel free to email me (SquawkWithMe at gmail dot com).


Everyone else, have this gif.


I was testing something I swear!
Logged

Mark Mayers
Level 10
*****



View Profile WWW
« Reply #65 on: March 18, 2016, 07:44:21 AM »

 Tears of Joy



Logged

Desolus Twitter: @DesolusDev Website: http://www.desolus.com DevLog: On TIG!
ActualDog
Level 3
***



View Profile WWW
« Reply #66 on: March 18, 2016, 02:18:42 PM »

Tears of Joy





Dude, you have no idea how much I want their secrets for balance (though I haven't looked at all so maybe there's a write-up somewhere). My best guess is it's something to do with forming triangles with the legs while balance is off. Maybe larger area triangles the worse balance is, increasingly smaller ones as it's coming back to acceptable bounds, and then correcting back to normal as soon as balance is back in an acceptable range. My joint setup makes it a little hard to use that strategy but I should probably play with a similar approach at some point.

On a related note, programming/designing big AI systems is very tiring and I keep taking breaks to tweak these guys' balance and rotation restore and stuff, because it's way more fun to do (though still important).

Tweaked a bunch of rotation restore stuff last night and then did some stair testing.


And then today I tweaked a bunch of stability steps stuff and found out that some of my math was completely nonsensical. I had written an angle subtraction function that was supposed to be returning both positives or negatives, but was always wrapping the angle to come back positive. Fixing that and tweaking stuff a bunch more got this system working worlds better than it had been.


Still not where I want it, but it's getting there!
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #67 on: March 21, 2016, 06:03:00 PM »

In the interest of having some sort of entry flow in by this Friday, I've started in on the opening.

I was able to re-use a ton of code from Animal Inspector (my previous game) for the dialog system, and I was able to salvage the bounce effect from some of my previous prototypes. I've never been able to recycle code this heavily before and it feels pretty great. I can get used to this!

But anyways, yeah, all of this is very WIP right now, and there's a great chance that it'll all change, but I want to get it blocked out in a way that doesn't kill my eyes. There's a good chance I'll end up working with an actual artist once things are further along, but I think it'll be nice to have even temporary stuff like this in for now.

The game's gonna start with you visiting the dog adoption center where you'll get to pick out your very own pup to take home and start raising. All the characters in the game will be literal dogs with dog names standing on their hind legs but they refer to themselves as people. You'll be raising non-literal dogs (the pets will NOT look like real dogs).



But yeah, this is the flow I'm working on this week. I'm gonna have a pile of pups drop out of a pipe in the ceiling and then you can choose one. I'll show the whole thing once it's working!
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #68 on: March 24, 2016, 03:47:32 PM »

Got sick so I've been operating at low frequency these past few days. I should be back to my normal output very soon, but here's an update anyways.

I added a system to my conversation controller for triggers and I'm working on a pipe system for the adoption center area that the dogs will fly through before being dropped into the pen.









I think I'm going to build an editor tool for these pipes. It's a pain to do by hand, I'll definitely need to modify them at some point, and I've never made an editor tool before so it'll be good to learn how.

Because I got sick I don't think I'm going to have this intro flow done by tomorrow like I planned, but hopefully soon!
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #69 on: March 25, 2016, 04:57:46 PM »

So it turns out that scripting editor tools in unity is super nice and a good thing to do. A little late to the party on this one, everyone's been saying it for a while, but I'm glad I finally got on board.

I built a little tool to make pipes with, so now something like this takes minutes instead of an hour and it's real easy to muck with.




The tool is hyper simple, it's basically just a list of directions.



It can theoretically work with any cube-shaped GameObjects that fit a few simple params. It auto-figures out pipe joints and works with the parent object rotating and all that good stuff.

For now these pipes are just for the intro, but this tool makes them super easy to make, so I'm thinking about other places I could potentially use them. Got some ideas but I don't want to say too much yet because it's all very tentative!
Logged

io3 creations
Level 10
*****



View Profile WWW
« Reply #70 on: March 26, 2016, 11:28:31 AM »

Let the Pet Evolution Revolution begin!  Grin   

Lol!  That part's really funny when so many pets fly up in the air. Wink


Have you thought of a title yet?

Maybe something along the lines of ...  Pet [R]Evolution  /  Pet R>Evolution
Logged

lobstersteve
Guest
« Reply #71 on: March 27, 2016, 04:50:07 AM »

That looks fun and is probably fun to make Smiley
Logged
ActualDog
Level 3
***



View Profile WWW
« Reply #72 on: March 27, 2016, 10:05:04 AM »

Have you thought of a title yet?

Maybe something along the lines of ...  Pet [R]Evolution  /  Pet R>Evolution

Not yet, but I really really need to come up with a working title. My goal right now is to come up with something by the time I get the first pass at this intro finished (counting sick delays that should be by Wednesday).

Let the Pet Evolution Revolution begin!  Grin   

Lol!  That part's really funny when so many pets fly up in the air. Wink


That looks fun and is probably fun to make Smiley

This project's had its fair share of frustrating moments (I was about a day away from quitting the project when I finally got a walk cycle working) but in general yeah, it's sometimes way too easy to get distracted while building this as you can probably tell by all the gifs. Hopefully that means it'll actually be fun to play too!

Speaking of gifs...


Logged

io3 creations
Level 10
*****



View Profile WWW
« Reply #73 on: March 27, 2016, 01:23:28 PM »

This project's had its fair share of frustrating moments (I was about a day away from quitting the project when I finally got a walk cycle working)
I know what you mean.  It can be very frustrating when things don't work out as you expect them.  Especially, when you are learning something new and may not even be familiar with every aspect.  I've been using Flash for a long time but started to recreate my game framework with Unity and found that (much like with Flash) I prefer to do things with code rather than with the Editor.  There were quite a few "is what I want even possible?" and wondered if I would have to do things with the "slow Editor approach" ...  Fortunately, there's a large community and lots of people had similar questions and so far I've been able to recreate everything I needed.  Only a few more things left. *fingers crossed*  Smiley
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #74 on: March 28, 2016, 07:02:04 PM »

I know what you mean.  It can be very frustrating when things don't work out as you expect them.  Especially, when you are learning something new and may not even be familiar with every aspect.  I've been using Flash for a long time but started to recreate my game framework with Unity and found that (much like with Flash) I prefer to do things with code rather than with the Editor.  There were quite a few "is what I want even possible?" and wondered if I would have to do things with the "slow Editor approach" ...  Fortunately, there's a large community and lots of people had similar questions and so far I've been able to recreate everything I needed.  Only a few more things left. *fingers crossed*  Smiley

Yep, just gotta keep truckin!


So uh, today something really "cool" happened. I've somehow managed to work on this game for about 2 months without realizing that Unity automatically figures in fixedDeltaTime when you call addForce/addTorque with the default forcemode. I figured this out because I needed to change the fixed timestep today to fix a collision issue and it broke everything in my game because I was trying to manually deal with fixedDeltaTime before calling addTorque. I'm really mad it took me this long to figure this out, and it was a pain to fix, but things should be back in a good place again.

That took up a bit too much time, but I also managed to make some more progress on the intro flow. Still on track for a finished first pass by the end of Wednesday, and then I want to move back to AI stuff for a bit.


Logged

Fuzzy Branch
Level 0
***



View Profile WWW
« Reply #75 on: March 29, 2016, 03:27:03 PM »

Looking good! The facial expression on the dogs seems appropriate for them while they're flying through the tunnels at the camera, ha!

Do you have a good idea of what direction you're aiming for with regards to the context and setting of the gameworld? Or are you thinking of keeping it pretty ambiguous and focusing on the game mechanics?
Logged

Tuba
Level 10
*****



View Profile WWW
« Reply #76 on: March 29, 2016, 03:31:07 PM »

I like where this is going
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #77 on: March 29, 2016, 07:42:40 PM »

Looking good! The facial expression on the dogs seems appropriate for them while they're flying through the tunnels at the camera, ha!

Do you have a good idea of what direction you're aiming for with regards to the context and setting of the gameworld? Or are you thinking of keeping it pretty ambiguous and focusing on the game mechanics?

Thank you, haha. Yeah their current expressions are made for moments like those.

I've started to think a lot about game world and context. I don't have too many concrete details to share yet, but I have a lot of ideas floating around and I'm hoping to be able to string them together into something fun and compelling. I'm really excited about this project and I want every aspect of it to shine!

I like where this is going

Thanks! Hope it continues to be interesting as it comes together more.


Today I got some basic scene transitions hooked up, did some more cutscene stuff that's not quite ready to show yet, and did some work trying to figure out how to map 2D mouse input to 3D motion so you can drag dogs around the screen without awkwardly using the scrollwheel to move them on the Z axis.

That last point turned out to be not easy. My current solution is to split movement in the y and z axes based off mouse y movement speed and to transfer movement from the y to z axis when a minimum's been reached (bottom of the stage). It works OK, but it's not shippable in the sense that I wouldn't feel comfortable handing the controls to a random person off the street. I'm also still not 100% sure I want to even let you pick up and manually move the dogs, so I might not have to solve this ever.

I also put in a title screen (as with everything, it's WIP) with what I'll potentially using as a working title. I might still change it before Wednesday (at which point I'll update the thread title) but I promised myself I'd have something by then. I think I like this as a working title, but I wish it had, like, one more syllable. I might try to come up with a word to put before it, idk.



ALSO, I've decided that I'd really really like to submit this to indiecade. The actual festival is like 6 months away and I feel very confident that I can get a fun build ready by then, but the deadline for submission is only about a month away and that's harder. I'm going to try and get a sort of sandbox version of the experience ready to submit. It'll be a rough deadline to hit, but we'll see where I can get!
Logged

tankorsmash
Level 1
*



View Profile
« Reply #78 on: March 29, 2016, 08:26:21 PM »

Are you going to do minigames, or like an open world breeding, where you need to have them live as long as they can, and deal with challenges like that?

As for a title, I think something that sounds snobby or refined would contrast pretty well. Like "Doggery, the Modern Dog" or something.
Logged

ActualDog
Level 3
***



View Profile WWW
« Reply #79 on: March 30, 2016, 04:36:14 PM »

Are you going to do minigames, or like an open world breeding, where you need to have them live as long as they can, and deal with challenges like that?

As for a title, I think something that sounds snobby or refined would contrast pretty well. Like "Doggery, the Modern Dog" or something.

My plan is to get the actual training mechanics farther along before I get too deep into designing events and activities, but I definitely want external stuff for you to do with your dogs. I want to give the player a reason to train/raise them in specific ways. I have some ideas I'm really excited about for all that, but I don't want to say too much until the dog AI and training stuff is further along, because that's gonna inform those designs a ton.

Also, thanks for the title thoughts. The more I think about "doggery" the less I like it, though I still don't have anything better in mind. Bleh.


In development news, I got my intro flow in. It's all very WIP and will probably completely change, but getting it blocked out got me to work on and think about a lot of higher level systems like scene transitions, dialog, UI, etc, and I'm glad I did it.





A lot of things there are ugly (sorry!), but it all gives me a nice base to polish away at. It also is really driving home for me just how little I know about 3D art. Still not at the point where I'd feel comfortable bringing anyone else on, but chances are I'll want to work with an artist further down the line.

With this intro blocked out, I feel relatively safe moving the thread progress to 10%, and I'm ready to dive right back into the depths of dog AI tomorrow!
Logged

Pages: 1 2 3 [4] 5 6 ... 37
Print
Jump to:  

Theme orange-lt created by panic