[UPDATED] ICECreatureControl v1.4.0 - creature AI for enemies, animals, monsters, zombies ...

Does anyone know how i can make the AI see the Player target while using additive scene setup, having the player in one scene(master scene), and the AI in another scene.

 

Do you mean you stream the active terrain, while your AI is on another terrain that is not yet streamed in? Or are you loading up your scenes into the game at the same time?

 

We had to sort this so I will share. It's a current limitation on ICE right now so you have to do a workaround (having to have everything that will ever interact with each other in the scene during setup and at launch). What we had to do was rely on TAGS for player targets. We also then had to create a dummy object in the hierarchy with that tag, and leave the GO enabled. Set up your interactors etc... to the tag. Now turn off the empty dummy object. Lastly, the objects with the TAG that are going to spawn into the scene and become targets, they need the Target script added to them. Now when they are added to the scene dynamically ICE will find them and add them as potential targets when the tag matches to the tag being sought after.

 

Its on Essentials Tab. Go into Default Behaviours. Enable IMPACT and set it to the BEHAVIOUR you want, then setup the behaviour. If you want a particle effect enable EFFECT, associate an effect to it and select TRG. That will ensure it plays every time the creature gets hit. However, as I mentioned, animation gets bugged when its time to DIE if IMPACT has an animation. There may be a workaround, I havent found it. I stuck to a particle effect and sound for feedback.

 

Thanks for the reply.
So for example, i leave an object with a TAG in the scene where everything is set up, and when i load that scene during runtime, i just transform that tag object/dummy to my player object(which is in another scene)? And they will find me?

 

Aww. Im having a strange issue with items being pickupable and suddenly not. I wonder if this is something to do with it. Items that are spawned are maybe considered different to the ones in the scene at the beginning. For example the drop item script actually destroys the item to drop and the instantiates another copy on the ground i think.

Following on some more testing it this seems to be the issue. Definate bug if when something is dropped it changes to an item that is not pickupable (presuming its not registered correctly in the creatureRegister. Add the "clone" to the name might be the issue so i will trim and see if it helps.

 

Give whatever you want to be found, tags. Doesn't matter where it is, instantiate it at run time, transition into a scene with it already there, it doesn't matter. Just tag everything you want found dynamically. Create the dummy object for it in the hierarchy so you can USE the tag, associate it in all your interactors(the creatures you want to find that tag, etc..), turn off the dummy object and you are good to go. This could carry through whatever scene changes you do, things can come and go, instantiate, get destroyed, it all works since you set up the pre-existing conditions initially via the tag. And again, dont forget, add the TARGET script to all the items that have been tagged so that when they enter the world the ICE manager is made aware of them and they are added to the hierarchy.

 

Indeed the object needs to exist in the hierarchy first. All my pickups work fine, but I created them first at ICE items and then also added the networking components to them. Make sure you enable detach on drop in order to be able remain after the body despawns so you can pick it up. You also have to set in 2 places - you create the inventory item I think in the area with Vitals or Essentials and stuff, then add the drop item in the behavior.

 

I'm not sure what's going on but I'm trying to use the Behaviour Layer Mask so that the AI will not damage each other. I just want the ICE Creature Melee Weapon to only damage the player. Without the Layer mask the AI can damage the player and other AI. I have assigned the player layer to player. Then I add a "player" layer mask to the Behavior Layer Mask. Now there is no damage being done to the player.

 

I have changed two things which seem to have fixed the pickup. One was to trim the name removing clone and second use the override function calling the item by name. This has made it rock steady and working everytime but now i have an issue of item not being destroyed but its far off now. I am also using my own pickup not the one on the Ice Player script as i need some extra functionality.

 

Hello! Any new updates? Im still designing my creatures so Im just checkin in to see what bugs have been fixed peace yall!

 

still waiting for the official. Hopefully it will be out by this Sunday!

 

Anyone using ICE with Invector Shooter template?

 

I'm hoping Pit will fix this but no Invector Shooter template doesn't work with Ice Creature Control which is a pity because Invector Shooter template is a great asset and I would love it if this would work with ICE. I have gotten RFPS and UFPS and even dismembering working with Ice. But on a side note Dismembering asset doesn't fully integrate with Ice I had to tweak a lot. But if Pit would do a proper integration with Dismember asset it would be greatly appreciated.

 

Thanks Wendy. Pit is obviously really busy as not getting any replies from Skype. But hopefully it'll be fixed in the next update - or I'll try and give it a go over the next few days based on his implementation and tweak it as required.

 

Ok first off thanks for the suggestions, i really appriciate it


So.. ive tried this for hours today, without sucess. Im not a pro in unity (yet), so excuse me if this is newbish


Scene 1:
This is where my player is, and always stays.
He has player TAG, target and Ice Player script. Also he has a target script (tried with and without it)
Since im developing in VR and using third person assets, i cannot transform my player to a hierarchy game object, because my player object is a child of other stuff needed.

Scene 2:

AI: Has everything set to TAG.

PlayerDummyObject: Has Player Tag and Target. Added to the CreatureRegister.

Tried the following:

1. Disable the PlayerDummyObject. Use the pool function in creatureregister to spawn a PlayerDummyObject clone, and set spawnpoint to TAG (Player).

• If i do this, the PlayerDummyObject will transform to the my player scene, the AI will walk against the DummyObject(clone) to attack it. If i delete the DummyObject during runTime, the AI will actually attack me, until the DummyObject respawns again. The Dummy Object will keep transforming to my player location after destroyed.
• I tried use a timed destroy script and add it to the DummyObject, which will destroy the DummyObject 5 seconds after the scene has opened. Well now the AIs will not attack me at all. It seems like i have to actually press the delete button during runtime for them to react to my player.

2. Just register the PlayerdummyObject in the creatureregister and disable it.

• When scene opens, and my player (which is in the other scene) is in the area. AIs dont react, move or do anything.

3. Register the PlayerDummyObject in the creature register and leave it enabled.

• AIs just walks towards the DummyObject and does not react to me at all.

+ the other 100 things i tried that did not work and i cannot remember

 

I ran into some significant issues when I upgraded ICE recently and through some debuging time I have found some code solutions to my problems. I figured I would post them here. I am going to post some code here and hopefully icetec won't mind.

This first part caused my waypoints to not be navigable. After deep diving I realised that I was generating Waypoints from Archimatix and they had wierd names like "Waypoint(Clone)(Clone)_corner_0_0" and "Waypoint(Clone)(Clone)_corner_0_1". This was fine in the previous version but in this one when it tries to select a TargetByName it cleans it with the following function.

Code (CSharp):

1.  
2. public static string CleanName( string _name )
3.         {
4.             if( string.IsNullOrEmpty( _name ) )
5.                 return "";
6.  
7.             int _clone_index = _name.IndexOf( "(Clone)" );
8.             if( _clone_index > 0 )
9.                 return _name.Substring( 0, _clone_index );
10.             else
11.                 return _name.Replace("(Clone)", "");
12.         }

This of course made all of my waypoints look like they had the same name "Waypoint" since it name.Substring right after the first "(Clone)". I made the following change (which I think might have been the intention anyway)

Code (CSharp):

1.  
2. public static string CleanName( string _name )
3.         {
4.             if( string.IsNullOrEmpty( _name ) )
5.                 return "";
6.  
7.             int _clone_index = _name.IndexOf( "(Clone)" );
8.             if( _clone_index > 0 )
9.                 return _name.Replace("(Clone)", "");
10.             return _name;
11.         }

The next issue I had was with getting my mobs to navigate a navmesh correctly. After the update they would head off towards identity no matter thier target. I use root motion for most movement and DEFAULT for the ViewDirection. The issue once I traced it down had to do with my navagent having a angular speed of 0. Now in previous versions this was fine but it looks like ICE kicks angular rotation over to the agent in this one. If its set to default it won't rotate AT ALL in ICE unless you set the navagents angular speed.

I attempted to do this but it was rather quirky to let navagent handle this. So I changed some code to get ICE to control rotation again in this intance. This isn't great but it does the job.
In ICE/IceCreatureControl/Scripts/Core/Objects/ice_CreatureMove.cs line 1788 I changed

Code (CSharp):

1.  
2.                 // CREATURE ROTATION BEGIN  
3.                 if( CurrentMove.ViewingDirection != ViewingDirectionType.DEFAULT )
4.                     Owner.transform.rotation = HandleStepRotation();
5.                 // CREATURE ROTATION END

to

Code (CSharp):

1. // CREATURE ROTATION BEGIN  
2.                 if (CurrentMove.ViewingDirection != ViewingDirectionType.DEFAULT)
3.                     Owner.transform.rotation = HandleStepRotation();
4.                 else
5.                 {
6.                     var r = Owner.transform.rotation;
7.                     Owner.transform.LookAt(_agent.steeringTarget);                  
8.                     Owner.transform.rotation = Quaternion.Slerp(r, Owner.transform.rotation, m_DesiredAngularVelocity.y * Time.deltaTime);
9.                 }
10.                 // CREATURE ROTATION END

This basically makes the mob look at the navagents current sterringTarget along the path. It still tries to use the AngularVelocity entered for the mobs behaviour. Which I "assume" is the intended behavior for ViewingDirectionType.DEFAULT if not that's what I want it to be for my version so I changed it .

Icetec if you have a better way to fix this stuff let me know. None of this is a dig against the product, it's a huge product with tons of lines of code. There are going to be bugs. Hopefully the community can fix minor stuff like this publicly pretty quickly. Hopefully someone finds this useful .

 

BTW.

I saw some people struggled with this earlier.

And oh man, after days working on a* and ICE, i have finaly made A* Pathfinding work with ICE(using the newest A* pathfinding) with root motion and local avoidance

If anyone still still struggles with this let me know.

Its just a few changes in the pathfinding adapter script.

 

Thanks for the share man, AAA

 

Hi! Can I check on the following:
(1) Is it still being updated/maintained?

(2) Does it work with Unity 2017.3 out of the box? To the extent that no compile errors or use of obselete APIs?

(3) Plan to support Unity Playable API?

(4) Is the source provided?

I'm currently reading through the 100+ page manual and I'm not under the illusion that it will be straightforward to use. Trying to understand where this product stands and where it'll be going. Can understand if author does not have the time to maintain it due to the cost/benefit tradeoff

 

Yep, that's what's going on I guess. I just wish the author of ICE would display the same commitment as the authors of some competing AI assets do... ICE is still the best, but I can't believe another half year has passed without an update.

 

Been around 4 months since last update.. won't be surprised if it's dragged on to half a year. I have some reservations the code might be in a poor condition (will explain partially on the lack of updates).

Emerald AI tends to get mentioned as alternative.. but it seems to be heavily combat focused. Closest I could find was ReGoap (available for free from github). Thinking of mixing it with a few other assets to get a full solution. But will be nice if Emerald AI or some other alternative can be used out of the box . Can suggest any? PM will be fine too. Thanks!

 

Not necessarily combat-oriented: I'm using it currently for kids playing on a playground. But it's really clunky. Animation cycles tend to be interrupted mid-sequence, NPC's circle around themselves instead of roaming freely. I guess that's the price of an easy setup procedure that is much simpler than in ICE. I'm trying to use Emerald in a commercial game, but it doesn't perform. So ICE is still the safer bet.

 

Its been over a month since I had any contact from Pit via Skype. So for now there is no sign of an update any time soon (until we hear otherwise). So we'll all need to help each other out for now.

 

In other assets, that would translate to being deprecated. Especially since 5 months went by already since the last update.

 

In addition, considering there has been at least 1 asset breaking bug (pooling) live for 5 months....

 

OK, how about setting up a Patreon for Pit? (I mean he could set it up for himself.) I would pay the guy monthly for even minor ICE updates.

 

I don't think its the monetary aspect of it, even though maybe the number of new sales may not be enough to support Pit long-term on this project, and so like anyone you have to find income elsewhere to pay the bills. I think he is just over worked at the moment and does his best to balance family, work and ICE CC development. So updates could take 6 months. I know that one of my products only produces two big updates a year for $199 subscription upgrade and they have a team. Considering that if I paid a developer to produce that big update it would costs me at least $1500-3000 per month full time (more or less).

So for the amount we paid for ICE CC its worth it. Ideally it might be better if bug fixes were rolled out quicker by producing point releases rather than having to wait for each one to be fixed and new features over a longer period of time. Easier said than done when you have to balance everything and as you are not in constant flow, it takes time to get back into working on each part and factoring what it may break elsewhere.

 

There is a thread about that subject, read it and you should better understand what is going on.
https://forum.unity.com/threads/asset-creators-not-supporting-their-work.518160/

Right, he has other priorities.
Ice plugin is so complex, even if it would go free, i'm not sure anyone would do bug fixes or bring new features (like SEGI plugin that is like an abandonned free plugin).
Anyway, i've moved to Emerald that provides very good AI, lot more simple and easy, and some minimal support from time to time, and because i just need basic common AI.
I can't wait months to get a simple bug fix with Ice plugin.
This is the problem with complex plugins , they ask author heavy support, it's not an easy task and you never know if the author will keep support or will have no more time some months after the plugin launch.

 

But from the end user perspective, if you a reliant on getting out a solid product you'd be understandably frustrated. If you could fix the bugs, it will solve all your issues, but there is also those that cannot, or would rather spend their time on other aspects of their game development rather than fix existing issues, but like any software it is rarely 100% bug free.

@Tethys - have you thought about using another pooling solution for the time being? I've never used ICECC pooling before, I use another products and it works fine.

 

I have both, just to see which turns out to be better for my needs. But ICE CC is so much more flexible in what you can do right now. Emerald is good, but is still lacking that Icing on the cake - which could take another 3-6 months to be where I could possibly use it.

If you are looking for quick and easy setup then Emerald is the way to go.

 

I suspect that is the case, and the recent 1 and 2 star reviews don't help... I think it led to a vicious circle where these complaining reviews (down to live bugs mostly) lead to ever diminishing sales which in turn discourage Pit to kill those very bugs and to post updates.
In part, Patreon was created precisely to address these kinds of situations.

 

Or at least upgrade cost to big releases (providing existing owners) upgrade and the majority of them would, which could sustain revenue for the year. An asset of this type may easily generate £2k ($3k) per month or more if it was marketed and maintained with small and regular point releases, with new feature releases as a chargeable upgrade. Maybe custom work could also be considered as a modular approach with a contract written to allow the developer to include these features into the core after a period of 6 months.

That's the trouble with one off sales, is that eventually that revenue will drop.. and you are still left supporting it, year after year for not a lot of revenue only the new sales that keep coming in that are based on new features or good reviews.

 

This is the problem, a popular plugin makes lot of money the first six months let's say, after a year the sales are not as high and the money income no more justify spending hundred hours on support and bug fixes or new updates.

 

You are right, take what suits your needs.
Ice plugin is super flexible, but new releases or small bugs can sometimes make your AI useless (with 1.4 i got ranged weapons issues and culling , pooling not working).
If i need to go advanced i go Behaviour Tree tools, Goap or FSM, so i keep maximum flexibility as i can do any AI i would like , while i got control on AI issues.
Anyway ICe Creature remains a great plugin, it's kind complex to hanlde, but it's the only one with so much flexibility while you don't to need to to code and use some BT or FSM tools.

 

Hey guys Need Some Help With Rag Doll . I have Enabled Corpse Option And Added A Ragdoll Prefab There Now When Creature Dies It Plays Death Animation Which Is 1.5 Sec Long So After Death Animation Creature is lying On the ground So i Have Set The Removing Delay Time To 1 Sec So The Ragdoll Spawn In Between The Death Animation , But The Ragdoll When Spawnned is Not Aligned With The Original Creature . So if i set removing Delay Time To like 5 second then at that time creature is lying on ground but when corpse spawn it spawns in standing position which gives an weird effect , How to fix this ?? any help

 

Should I buy this Asset ?

I bet like me other people visit this forum to help make that descision.
There have been some bad reviews lately but I feel you have to read between the lines especially with assets like this.


Out of all the categories on the asset store “Scripting AI” must be the most difficult to create maintain and support, certainly compared to an asset like ground textures or a low poly model.


The 5 star rating system is not fair because its very easy to get a 5 star rating for a low poly kitten model where the support consists of replying once per year to the “has it got whiskers question”

You’re always going to get people expecting assets like this to do more
or do less
or be easier
or more comprehensive
and some who expect it to make the whole game for them.

So here is how I’m arriving at my descisoin:

Does the asset seem to do what I want?
Is the author still around to help me if I get stuck?

If the answer is Yes then it’s a no brainer I shall buy the asset.
Worst case scenario is: I helped support someone who is trying to make it easier for me to make my game.
Best case scenario is the asset saves me a massive amount of time trying to Write a AI system.

Should I buy this Asset ?

 

I think you're kind of comparing apples to oranges. Of course 3d models require different level of support than scripting/ editor extensions categories. Still you can have complex modular 3d models packages which could get quite involved to use and require a lot more help from the authors than a low poly kitten.

Regarding your questions;

Thats not a very good question without saying what it is that you want. Do you want simple ai for wildlife that minds its own business and sometimes chases the player? Do you want human ai that has its daily routines? Do you want ai for birds that are just in the background?...

The author is still around but not very active. I believe there will be updates in the future, but hard to say when and how often. I wouldnt expect fast help if you encountered any issues.

Probably. Hard to say since you never mentioned what it is that you really want. ICE works for the most parts, but sometimes if something is not working as expected its hard to say if there is a problem with the way you set it up or if its a bug. The more use cases you will try to cover with ICE the more noticable this will be.

 

Other questions;

Are you experienced enough to write your own AI
How many hours vs your time vs your money would it take to create an AI system?


If you can use an existing one and it will do for now until either you get better at coding, or a better solution comes along, or you've made your money back on the asset and more and now you are writing your next game.

 

The real question here is: how much time you will spend waiting for the author to support vs. how much time you will create the same (or proper) results with other, actively supported solutions?
I have read through the entire thread, unfortunately I have to say, if I haven't own this asset, I wouldn't buy it. The author tends to disappear for quite a lot (for months!) without a word.
He logs in to the forum but can't post a "hello, I'm busy, I'll be back on 21st of March in 2034" message.

Yes, I'm a developer, I know how it works, but I don't see why the author does not come around once in every two weeks and spend a day to answer the questions. That's all. Wouldn't be a big deal if he actually care about his community.

And my sincere gratitude to you guys, who began to support this product. For free.

 

I agree with you. If he doesn't have time, it should be made free - the more people using it the more the community (may) improve on it and share what fixes or enhancements they have.

 

I think the plugin doesn't make enough money that would justify putting an entire day that could be a paid one.
Anyway, it is as it is, i got another plugin about Tps and Goap system, i never got an answer, one year later the author tells it has a new update and it's improved, he respond to some questions one day after new update and he disappears again lol

It's not sure community support will happen ? Look at Segi plugin it's abandonned and someone is trying to make a new plugin and better. So make it free perhaps won't change anything.

 

On the other hand, take a look at the Community Ocean Shader. If there are a couple people, who likes the project and willing to use it, support it and develop it, it's all set.

 

This. I think that the concept and implementation of ICE comes near what we could call genius. I mean no other AI asset comes close. But clearly this genius has moved on, leaving even the manual and the tooltips unfinished.
If Patreon-based funding won't motivate him now, he could make ICE open source, or at least drastically reduce its price so it could finally reach wider audiences.

 

It's not so much supported, someone reported some foam rendering issue one moth ago, and still no answer or correction is proposed . Perhaps i'm wrong.

 

It's a shame as a few more solid updates could have been it. I know of another AI system that was given to another developer, the original author had a small cut and they are now pushing updates out fairly regularly. That could be an option providing someone pushes it forward and not let it sit on the shelf, like it is now. I would welcome if it became free or MIT licence.

 

Well, it's a free package, you get what you pay for. But you have all the source so you can fix it yourself. Or you can wait for the community to fix it. Now, let's check the same with ICE... How many bugs are reported for at least six months and how many are fixed? And how many you can fix?

 

Or you can fix

 

I'm sorry that I could not answer your requests earlier but I had a serious accident in January and was in the hospital. I've been home for a couple of days now, but I'm still under medical treatment and it will take some time to get fully back to work. But however, I will try to answer the open requests and publish the announced update in the next few days.

btw. the update was already nearly finished before I had the accident, so it should not take so much time to publish it.

Sincerely yours,
Pit

 

@icetec, we are deeply sorry to hear about your accident, and wish you a swift recovery.

In the meantime, please consider taking advantage of Patreon for the further funding/development of ICE (as discussed in above comments). I'm sure plenty of us would gladly contribute on a monthly basis.