Search Unity

  1. Tune in to get all the latest Unity news live from Berlin on June 19, 6pm CEST. Set up a YouTube reminder!
    Dismiss Notice
  2. Unity 2018.1 has arrived! Read about it here
    Dismiss Notice
  3. Scriptable Render Pipeline improvements, Texture Mipmap Streaming, and more! Check out what we have in store for you in the 2018.2 Beta.
    Dismiss Notice
  4. Improve your Unity skills with a certified instructor in a private, interactive classroom. Learn more.
    Dismiss Notice
  5. ARCore is out of developer preview! Read about it here.
    Dismiss Notice
  6. Magic Leap’s Lumin SDK Technical Preview for Unity lets you get started creating content for Magic Leap One™. Find more information on our blog!
    Dismiss Notice
  7. Want to see the most recent patch releases? Take a peek at the patch release page.
    Dismiss Notice

SwarmAI 3D PathFinding System - Should I sell asset?

Discussion in 'Assets and Asset Store' started by FirstTimeCreator, Jun 7, 2018.

  1. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    Hell Guys,

    I have been working on a 3d pathfinding system and it is getting ever closer to a production quality asset, What are your thoughts on making a sell-able version on the asset store?

    I am also building 2 other things which will be the core foundation for the space sim im building which is:

    1. Squad Formations, Squad Leader AI
    2. Team Leader AI
    3. Weapons systems
    4. Mock battle between 2 teams.

    I don't know if it would be worth it to invest the time into creating an asset for the store or just continue building the game.



    Improved steering in latest update:

     
  2. EternalAmbiguity

    EternalAmbiguity

    Joined:
    Dec 27, 2014
    Posts:
    2,115
    Keep in mind that most people with assets like this, actual systems not just models, seem to not actually develop games, but spend most of their time improving the asset, fixing bugs, providing support. So unless you're okay with putting that kind of time aside on a regular basis, not just as a one-time effort to publish the asset, you might want to avoid that.

    Maybe @TonyLi can chime in on this.

    Outside of that though it seems super interesting.
     
    TeagansDad, elmar1028 and TonyLi like this.
  3. TonyLi

    TonyLi

    Joined:
    Apr 10, 2012
    Posts:
    7,919
    Yes, EternalAmbiguity is absolutely right. Time spent fixing bugs, making improvements, updating code to handle API changes in new versions of Unity, answering questions, managing publisher stuff, setting up a website, writing documentation, and recording video tutorials (and customers will ask for video tutorials) is time that you won't be able to spend on your game. However, being able to help others make their games by providing a good asset is rewarding itself. Be prepared for that to be your only reward. Selling and supporting an asset on the store isn't a golden ticket, especially for a relatively niche (albeit very interesting) asset like 3D pathfinding. I'm sure there will be a number of people who would find invaluable. But most people are working on RPGs and platformers, and they have different pathfinding needs. I'm not trying to discourage you, though. It's really rewarding in its own way to see your asset being used in other devs' games.
     
  4. snacktime

    snacktime

    Joined:
    Apr 15, 2013
    Posts:
    1,440
    IMO you need to either go the complete game starter kit approach, or go the opposite direction and do one thing really well.

    I would be extremely suspect of an asset that tried do to multiple things where one of those is pathfinding. Even more so when it's 3d pathfinding. There are some non trivial problems to solve there for it to be any good.

    What you should do IMO if you want more valuable feedback, is give more details on the stuff that matters most. What is your approach to the pathfinding and what are it's limitations. How is it better then other solutions or what we could do on our own.

    Same for the ai, what is the basic approach there. Is it modular/extensible, etc..

    Does it use the new job system? I think that is key for new technical assets doing this kind of thing.

    Bottom line is the barrier to entry on technical assets is going up. Which is why I say complete game starter kit, or do one thing better then anyone else, or just go make your game.
     
  5. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    So far there are no limitation other than a maximum velocity of 500, that may soon change though as im probably going to expand the grid and casts to account for extremy fast velocities, most ships won't move that fast in the game world though unless in hyperspace and the collision avoidance is disabled while in hyperspace flight so...

    My approach is i use a proceedural grid around the ship, i linecast and sphere cast efficiently to achieve up to 200 ships in a scene or more under 3MS response time on i7.

    It's really difficult to explain it fully in a forum post because it is really complicated. Right now im working on the "rubberbanding" when the ships are turning toward thier target direction.

    Even in an asset, if someone wanted to learn "how" this works they would have to spend alot of hours studying it.

    Your right though, im more interested in building a game.

    My next video will be awesome, ive made SO many improvements, i have all 200 ships hyper jumping to a destination 200K meters away, as well as massively improved steering behavior with very slight rubber banding im still working on fixing.

    Ill have the next video out by monday with about 30 hours + of improvements and additions.

    I spent 2 years learning how to build this pathfinding system by the way, through experimentation ect. this is my 4th total re-write (re-design) of the system (and my final).

    Oh and floating Origin is active and enabled in my next video as well :). I don't think ill have time to do LOD systems and a que system for the computer voice i added to the cockpit of the ship that says right now things like "Waypoint reached, Collision Alert, Prepare to enter hyperspace" Ect.

    Ill post the next vid here when i release it.
     
  6. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    It's all modular, expandable. Each component has it's own class and the code is broken down into the smallest possible components.

    The AI will be integrated in a way that you could swap it with a different AI simply by changing a class.
    Does not use new jobs system, no need. I will be using the HyperThreading processor capability for "weapons systems" later, if that is what you are referring too.

    It will be pretty much required with this many ships.

    Yah, ive given it thought and i don't think ill do an asset because, its really hard to find the time to build the game i want to build, if i did asset store i don't want to get tied down in that and not produce an epic indie space shooter that ive always dreamed of building and does not exist.

    one with EPIC space battles with AI pilots that can navigate around complex and moving geometry and is good enough to chase the player around said geometry.

    Not only that but the scene area is 2 million meters, and ships are able to hyper jump to other planets in the scene, i actually have that working now and it will be in my next video. It's pretty F'n Epic!
     
  7. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    Well I have them all turning fluidly without rubberbanding after hitting their target direction. This presented a new set of issues which im going to fix by implementing a "hard turn" system based upon the distance from the object. So if the ship gets close to an object and won't make the turn using it's current torque forces, it will know to boost the rotational forces.

    It is very difficult to account for all of the scenarios given that objects are not static, which would be a lot easier, since they are moving and rotating that presents and even bigger challenge lol.

    Still i would imagine that this will get my crash rate down, building a "perfect" system that can run 200 ships without adding additional spherecasting systems would be difficult.

    So im shooting for a 90 to 95% success rate of ships within say a 10 minute period of them flying around objects. Eventually some will get caught, bump into each other and then crash and there is really not much i can do about it because of when they cluster together, im not sure how to really "break them up" if they are in a confined space say between 2 moving asteroids.

    Anyway i have been hard at it for a while now, im so close to making a video that will blow your minds :) so close.

    In my demonstration ships cluster together, but what i have planned for the actual game that will use this system they won't all be clustered together is large groups as they will be in different areas in different battles traveling around the solar system. They will travel in squads of 6 or less and fleets of up 25 to 50 ships.

    I have a system to detect a crash basically, i was thinking of implementing logic to get the ships to "reverse" and turn around. Or it could actually be interesting to leave it like that and allow npc's to crash and have some nice pretty explosions going off for the player to enjoy :)...
     
    Last edited: Jun 10, 2018
  8. newjerseyrunner

    newjerseyrunner

    Joined:
    Jul 20, 2017
    Posts:
    193
    I’m confused, you said the AI works with a mesh that surrounds each ship? What if the destination is beyond that? Does it have an A* routine? Like can I put it outside of a huge maze and tell it to reach the center?
     
  9. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    No my system does not use A*, it is a custom designed system, the only one of it's kind i believe as i came up with it myself and there is no examples of tutorials of a system like mine on the internet.

    Yes, the ships will reach the center of the maze, even if the maze is "moving" around. There is no mesh around the ship, there is a procedural generated Grid that i created that has a specific shape. The system uses the grid to check for possible course corrections, it also looks for the shortest available path towards it's intended destination first as well to that it attempts to take the shortest path toward the target if it is available.

    With my system the distance between the ship and it's target does not matter, nor does the environmental geometry matter as long as the ship can fit through and around the environmental objects it will make it to it's target.

    I will release another video later today whether im ready or not to show you guys the progress i have made which is alot, since ive put 40 more hours in or so.

    The only thing im working on now is logic for getting ships unstuck that crash, because of the number of ships flying in close proximity crashes will occur even with the best path finding system because they bumb into each other if there are 20 ships flying right next to eachother around an object.

    so im going to do is get the ships that crash into an object to back up and turn around. I'm also implementing a "hard turn" system so if a ship gets to close to an object and can't turn away with the base torque force it will amplify giving the ship another chance to avoid crashing.

    After that ill look into implementing a "roll" system, that won't be in my next vid though.
     
    XCPU and newjerseyrunner like this.
  10. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
  11. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    Lol ive done it, i put like 45 hours in, in 2 days.

    No more physics rubber banding, ships yaw and pitch now. done alot of work, im uploading a 4gig video now which will probably take 15 hours to upload. I'm just excited to share it.

    I got 304 ships to run at 200FPS average 275 Peak, Stable, no memory leaks, crashes or issues what so ever.

    My pathfinding system is just amazing now, the only one of it's kind, you will see in the coming video.

    Took me 2 years on this journey learning unity and studying everything i could get my hands on with pathfinding systems and building countless other prototypes.. trying and failing and trying again, learning and improving with each iteration.

    Mastering floating origin, editor extensions... Believe me when i say there is alot more too this than simply rotating the scene around the ship, in order to make it work. alot more...

    But once you master floating origin and work arounds to make things work properly, well you have what you need to build a REAL space game.

    One without boundaries. With REAL "warp drive" where the ships actually travel and light speeds to there destination.. No fake warp jumps.

    And with the CPU threading system in 2018, I will be able to push more performance out of it and make the weapons systems work well because, there will be alot of gun fire coming off capital ships and alot from fighters as well.

    When im done, it will look like a F'n Apocalypse in space in the most epic space battle you have ever witnessed in a video game.

    And i have started on an epic, all original story line as well.

    Once my prototype reaches a certain point, after weapons systems, mock battles and 2 additional AI systems (squad commander + team/fleet commander)

    I will start reaching out for modelers that want a share of the game to build ships for me. Because I don't have the money to pay them I'm definitely willing to share ownership of the game to a modeler as determined as i am to break out into the gaming industry and create an awesome space game.
     
    jackmememe and XCPU like this.
  12. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404


    Galactic Armada: The Fall of Humanity
    The first game in the series will be a single player story driven space action combat game
    where you play the role of a squad captain. The game will feature some open world game play as well as some linear story line
    arcs with a flow similar to that of the space sim classic "FreeSpace 2". My intended target for total play time before the game ends is at least 10 hours.
    Because I am a solo developer doing pretty much everything, the game will mostly feature text based GUI prompts for the story in which the player must read, with some
    voice over's that i get my friends and family to help with.
    In the first game in the series you will be introduced to the
    "Galaxy", 400 years in the future in the year 2,518. In this distant future, mankind has expanded into space, constructed colonies on other worlds and established a
    militarized presence in space. For 100 years mankind reveled in their accomplishment, arrogantly believing in their technological hegemony over
    the few alien races in which contact has been established.
    And then.... They arrived, with fleets so large, not even light could pass their awe inspiring presence. That is when everything changed.

    -----------------------------------------------------

    Galactic Armada: Fire,Fury & Conquest (expansion)
    If the first game in the series is successful, the second game will be an expansion which will add a Conquest system to the game.
    This system will feature Sectors which contain Solar systems contained in a Galaxy Map similiar to the X3 game series.
    Players will be able to construct new ships and give orders to fleet commander (AI's). Form fleets and capture sectors for resources.
    The game's conquest will be simple based upon a "Risk" model where sectors produce a certain amount of resources & ships each year, which the
    player will allocate for building new ships and researching ship upgrades.

    My goal is to create what X3 was not, a conquest game, but without the trading. Exploring will be a possible addition, however
    the point of the game will be strategy,conquest & fleet command. Exploration type missions could be implemented to grant players "bonuses" through artifacts.

    Additionally, the fleet engagements in Galactic Armada will be much larger, with many more ships than what was possible in X3 using mods to drop ships in. The x3 engine and software design simply could not handle anything
    past 100 or so ships without completely bogging down the engine.

    My system is being designed to run up to 200 smoothly on an i5 CPU, 300 on an i7 and up to 500 (within a single scene,spread across a solar system with active collision avoidance, path finding and multiple AI systems working in concert) on the latest generation Intel CPU's.
    The game will be designed to allow (scaling number of ships) options based upon the CPU the game is running on.

    ------------------------------------------------------

    The game is ambitious for a solo developer and the software design to run it will must be a technological marvel to behold as well as a
    true demonstration of the power of the Unity Game Engine. (Amazing). I think it would also be the first "true" space game built with unity as all i could find was some very basic prototypes.

    Building this game has been my life long dream, my passion. I was a programmer for 15 years prior to starting to learn game development over 2 years ago.
    In the past 2 years I have not given up, I have banged my head against the wall a few times but I have persevered and I have finally achieved
    the first and most difficult component required to run Galactic Armada, the Collision avoidance path finding system, capable as well as efficient enough to
    run hundreds of ships with enough CPU left for weapons systems and AI.

    So tell me, Who is ready for a new Space Sim Classic? Who wants to relive the days when space sims were new and refreshing and pushing the
    boundaries of technology as well as "fun" with a good, interesting original story. I know I do, and I know it won't happen unless I build it.
    The time of waiting is over. The time of creation has begun.
     
    Last edited by a moderator: Jun 17, 2018 at 6:29 AM
  13. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    1,889
    Congratulations on your progress so far with this. You are off to a solid start.

    You asked in the original post about whether this would be a good idea for the asset store. There is definitely a demand for a 3D path finding solution for space games on the asset store. The challenge is that each developer is building something a little different, so it would probably be tricky to build and support a one size fits all 3D path finding solution. But I do think there is demand for this. Even if it would not work in every 3D space game, it is possible that quite a few people might purchase the asset just to see how it was implemented.
     
  14. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    Mine would work in any space game, actually. I really don't see a reason why it could not be implemented into any space game as it has a API bridge controller for an AI class to intercept and modify behaviors. Technically you could write your own AI and still use my pathfinding system with it pretty easily.

    The thing is though, i spent so much time trying to figure out how to make this work lol.. would i really want to give it away. When it's all said and done im talking thousands of hours when you include the AI I am building for it which will have the capability to perform maneuvers and react to battlefield conditions.

    This isn't Star Citizen (Cough), my AI will actually do more than fly around in circles.

    If i put mine on the asset store it would be the best one on the asset store, I have seen them all and as far as i know there are none that can navigate complex MOVING geometry using 100% physics.

    It was really just a though because I need money for modelers... I've decided im going to build the entire game and use placeholders for where artwork will be and I will use sketchup models until I can get actual models for the game.

    I will be seeking investment and at a certain point in the prototype development I will start pitching it to medium + large game development studios that have modelers in house.

    I will make a "ready made" game ready for models, art and voice overs, complete with story GUI and playable. Who the hell would refuse that deal...they would have to be stupid.

    This game will be the most advanced space simulation combat game ever created. I am taking skills I deployed in other fields, namely trend analysis and prediction systems and constructing a precognitive reactive AI which will be a lot more than a simple "weak" AI.

    The AI will be able to formulate new strategies and react to battlefield conditions. I'm talking about the Team Leader AI, which will give orders to Squad leaders, Fleet Captains and Fleet Admiral AI's.

    At the core of the game though, besides AI, is the pathfinding system and collision avoidance which needs to work extremely well as well as be efficient enough to handle up to 300 ships in a scene (which I have achieved) still leaving room for weapons systems and CPU for AI systems.

    It's actually so good, I'm achieving that response rate using a single core and once i migrate to 2018, I will be able to push more performance and definitely intend to run my AI's in parallel as well as a few systems I have implemented right now.

    It's a shame you can't thread physics processes though, the possibilities would be unimaginable if you could.

    Speaking of the CA system, im actually about to make a big improvement,massive improvement. I am replacing sphere casts with a procedural line cast "fork" system I am designing. I am hoping that it nearly doubles performance, might even allow up to 500 ships in scene.

    But... with so many ships in scene load times are extended IF there are that many starting out when the player loads the scene. So, If i go that "massive" i may end up making the entire game run in a single scene and be data driven as far as object placements. Meaning the entire game will run in a single scene and objects will load and deload based upon the player position, I'm using a floating origin system where the player never leaves 0,0,0, even while appearing to "move". There are alot of challenges no doubt, but the "hard" one's ive already worked out :)

    One thing about a collision avoidance pathfinding system, its not possible to make a "perfect" system, especially given that I have to balance it with the need for extreme efficiency to allow large numbers of ships in scene. But it IS possible to achieve a high success rate of ships making it through complex obstacle courses, im shooting for 90% or higher.

    I'm actually uploading a video now where they all make it because the ones that crash, back up turn around and try again lol. I implemented a system to get the crashed ships put it in reverse, back off the object and then look for a path and make the attempt again. In the game though there will be a damage system so depending on how fast the NPC ship is going it may just explode if it does'nt make it. I thought it would be cool to see the stupid pilots not making it, actually seemed realistic lol. And since there will be so many ships fighting the player. So, I intened to make the weakpoint of my system the "imperfect" part a part of the game :)

    When i am done, the AI should be able to chase players around moving objects, even through holes in asteroids with a high success rate.

    This your game?
    https://store.steampowered.com/app/637410/Disputed_Space/
     
    Last edited: Jun 17, 2018 at 7:40 AM
  15. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    Oh man, it's quite complicated, they would need to be a solid unity programmer to understand it and quite a few hours of time on their hands. It's growing ever more complex as well as I improve it, optimize it, add behavior methods.

    Which is exactly why, no, im not going to sell it. Because i know i will be flooded with so many questions....which i don't have time for really, since im working around the clock from wake to sleep to get Galactic Aramda off the ground, to turn a prototype into a game. A game a publisher would pick up and fund.

    No one is going to make a tutorial for something that involves thousands of lines of code and extremely deep logic. I could however point someone in the right direction lol, with a enough time, dedication and sheer determination they could build their own...

    One thing I have learned in the process of learning how to do this is there is a reason why this does not exist, not publicly (no tutorials) on how to make something like this and i have never seen it work in a space game on this scale and Ive played them all from X3, freelancer, Evochron. I love space games which is why I HAVE to do this. Someone has to make a game that will advance the genre and Star Citizen isn't it. So disappointing. I want to bring back the days of the fun classic space sims, but bigger, badder, more beautiful and smarter.
     
    Last edited: Jun 17, 2018 at 8:34 AM
  16. HeadClot88

    HeadClot88

    Joined:
    Jul 3, 2012
    Posts:
    627
    I would buy this if it was put on the asset store. :)
     
  17. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    Oh man I just had a great idea, instead of a fork, im going to spin the line casts in a circle radius, a single line cast to replace spherecasts for detecting surface geometry.

    oh boy i'ts going to work to i know it. Watch 400FPS with 300 ships in next video.

    Anyone want to bet 10$ lol. J/K.
     
  18. XCPU

    XCPU

    Joined:
    Nov 5, 2017
    Posts:
    41
    Looks interesting.
     
  19. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    O M G

    IT WORKED!!!!!!

    ZERO sphere casts and it works BETTER than when i was usng the sphere casts!!!

    500 ships here we go 4MS response time or under!! (these are unoptimized models, so there will be somewhat of a performance hit from them being sketchup models)

    I did remove all backface materials however on all the models.

    I will be posting a new video later today AFTER the one ive already been uploading for the past 8 hours... my upload speed sucks.

    I will also show you a development view, so you can see what it is doing, i will let you see my "visual representation debugger".
     
  20. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    I guarantee you after you see what this does in my visual representation you will believe no one has a system like mine, NO ONE. :)

    2 Years in the making and I've achieved the seemingly impossible, at last.

    There will truly be EPIC ARMADAS in Galactic Armada. The game will live up to it's name. Imagine 500 ships going at it in the epic climactic ending of the game, or 200 vs 200, with frigates, carriers, fighters, gunships, support ships, space stations in the middle.. gun fire everywhere. Oh yes I can't wait to start on weapons systems...

    but first, i have more work to do to perfect the collision avoidance steering behaviors for example im about to add a roll system so that ships roll into their turn as well as pitch and yaw so they appear more "fluid" and realistic. In the prev videos there is no roll, it will look and perform better with it.

    After that im doing the player flight controller so i can fly the ship around and then.... weapons systems starting with turrets, a basic friend foe system, targeting system and "chase" system which will be the base architecture setup that my AI's will plug into later.

    Basically I'll be adding whatever I can think of that I can get ships to do as "commands" that they will receive from their AI overlords.

    I'm pretty creative so im pretty sure I can put EgoSoft to shame :) as far as what ships in the X series could do.
     
    Last edited: Jun 17, 2018 at 12:22 PM
  21. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404


    I was on the fence about publishing this vid since ive since increased the performance and collision avoidance has been improved drastically. I uploaded it so here you go.
     
    HeadClot88 likes this.
  22. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    1,889
    Yes. That is my game. I released it last year on Steam. It is a 6DoF arcade shooter. Think of it like the old Descent games, but in space instead of inside a mine. It is a game about strafing and blasting.

    I developed a lot of highly optimized AI for my game. I absolutely love digging deep into AI and building layers of AI to operate epic scale space battles. The feedback has been very mixed for it, though. Some players really loved it, but a lot of players would have preferred different design choices. A lot of people could not see everything that was going on, and they were unable to appreciate all of the procedural splendor.

    I did get a lot of extremely valuable feedback, both for improving aspects of the existing game and for better designing future space games. My next space game will fly more like a flight simulator and it will be extremely story driven. My next space game will actually have less of the procedural solutions and more completely scripted sequences.
     
  23. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404

    Oh nice.

    Well have you ever played the X series? I've always wanted something similar to the X series but 100% focused on combat and fleet combat management.

    X games were fun in their time, but the problem is they tried to do too much. Trying to do trading, combat, building all in one game makes it too where you don't do anything all that well.

    The interfaces were far to complex and the learning curve too great for example.

    My game will have a sector based "galaxy map" similar to X3.
    The initial launch of this game won't have the conquest elements though, it will have the building blocks for it and be designed for it. In order to get the game out the door, or at least a fully functional prototype playable game lacking models art work and good sound effects, I have to focus on just a single player campaign, while implementing the ability to expand into the conquest version.

    What is Galactic Armada exactly?
    A Fleet Combat Simulator + Space Action Combat Flight Simulator and eventually a 4x space strategy game (if the single player campaign version does well).

    In the expansion after the single player campaign, the player will be able to take on the roles of the AI EMPIRE leader and perform all of the actions and commands the AI can through a GUI.

    I won't be working on the Conquest AI though in this first game. Only squad leader and fleet leader AI's.

    The player will use the same systems the AI's have access too, since im bulding the AI's first and the GUI's after for the player.

    The player will be able to control and give orders to the fleet in the single player campaign as well as fly various ships from fighters, gunships and capital ships. Capital ships though will be played from an exterior camera perspective. I have no plans for doing cockpits or interiors for capital ships. I do not have the resources for that.

    To make this game possible I have to stay RAZOR focused on what I CAN accomplish, working my but off... 100 hours a week.
     
  24. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    And then after i have a fully playable GAME... I will beg, borrow and do what ever I have too, to get the rest of the content, art work, models. ect.

    Everything at that point will be "drop in" so all i will need is either money or partners and I am willing to give % ownership of the game if I have too.

    The vision for my game is awesome and I know that I can pull it off so I'm not that worried about getting funding after people are able to play the game and see what IT is.

    unlike the vapor ware out there, I have full plans to distribute limited versions of the prototype to the masses, a playable scenario of fleet combat.. for free.

    All of the models though will be from sketchup and the models won't be that great but.. the programming will be. :)

    I'm also a full stack web developer and a SEM specialist, so I can build the website and do the marketing.. with what little I have. Finding the time though... to do the website, while working on the game.. yah I have no plans for a life whatsoever for the next year.
     
  25. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    MultiPlayer for the conquest expansion is a real possibility, however my only concern there is tracking the weapons systems over net. Tracking ships shouldn't be that big of a deal.

    I intend on design a raycast based system that simulates bullets and particles to represent bullets as to not have 50,000 colliders and rigid bodies flying around. If not particles a transform based bullet system with raycasts and a simulated bullet collision, at least for the turrets on capital ships.
     
  26. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    1,889
    No, I have not played the X series games. I played a lot of Master of Orion, which has some similarities. I definitely agree that doing too many different things can lead to game mechanics that feel weak, especially when it comes to combat.
     
    HeadClot88 likes this.
  27. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    1,889
    Yeah, you definitely cannot track each projectile in real time over the network. To do multiplayer in a massive space battle, you need to build a deterministic simulation. Then queue some information about what individual ships and turrets plan to do in the near future. Then periodically burst those blobs of intent to the clients. For example, a turret might plan to target a specific fighter ship and then fire a bunch of shots with random delays between shots. Instead of sending network data about each shot, generate a blob with the information for the target selection and delays for the next ten shots. Send that blob of data to the clients and let the clients play that data in their own simulations.
     
    HeadClot88 likes this.
  28. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    I feel like there hasn't been a fun space game since freelancer, even though the AI wasn't that smart it was a fun game.
    I feel like no one is really developing AI for space sims, not really.

    What i mean is I would be willing to bet you that in X5 game, the ships fly around in circles and have no plan, no strategy.

    AI is just weak in every space sim game that exists. Why?

    I don't understand why it has to suck so bad. Is it because most space sim games try to do too many things?

    I'm on a mission to prove AI in a space sim game does not have to "suck" by default.

    From my experience playing just about every space sim that exists, they get it working and do nothing beyond getting it to chase and fly around in circles, there is no real intelligence its not an actual AI. AI implies programmed intelligence and decision making beyond "run, chase, shoot".
     
  29. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    My AI will have Kamikaze, a wide range of pre-programmed Maneuvers for individual ships and squad formations. Within a squad, pilots will communicate with each other and tactics will be adjusted by the state of the group.

    Say for example half of the ships are damaged in the squad, or there is just one squad member left, that pilot may decide to overload it's weapons at the risk of destroying its own ship or decided to attempt to ram the player.

    It might decide to run and hide behind some asteroids and try to ambush you, force you to chase. I can think of alot of gameplay mechanics I can design that have not existed in a space sim.

    :)

    In fact it is my intention to build those mechanics that have been lacking and never before seen in a space sim, especially a combat space sim, which is why the collision avoidance pathfinding system is SOOO important for my game.
     
  30. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    1,889
    Honestly, users don't really want layers of amazing AI. They want a story, voice overs, and very specific mission goals. They want extremely clear individual tasks, and they want the game to basically wait for them to finish each clear task. Even if they say they want awesome AI, most users really don't. Unless users can see and directly interact with each AI system, those users won't appreciate any of the AI. All of the AI stuff that you and I get excited about is largely ignored by users.
     
  31. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    That's because a game like Galactic Armada does not exist yet. This game will have every element you just listed and a good AI is a core element to it working.

    It's a fleet combat simulator not a spray and pray kind of game. You will be giving orders to other ships and expecting them to carry those orders out in an intelligent way.

    In this game the AI will be the heart and soul of what makes it a great game, and the second thing will be the combat the player interacts with by flying around and taking part in the combat.

    I have a plan for a VERY intuitive control system for the player to be able to fly and engage in combat AND order the fleet's ships to carry out tasks.

    And the bonus.. it will work in VR with oculus rift.

    I'm planning for VR, do you see the way the offsceen indicator works, so the player can move there head around and see where ships are easily.

    Soon ill show you what im working on for the VR interface, i actually have something in place just havent shown it in video yet.

    The VR interface will work both on PC with mouse and with a VR controller.
     
  32. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    That is exactly what they will be doing in Galactic Armada!! its the only other thing they can do besides fly ships and engage in combat.
     
  33. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    1,889
    It sounds exciting. Good luck with it.
     
  34. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    Oh boy, ladies and gents.

    I have done it. I have created PERFECT pilots.

    AI pilots that can navigate ANYTHING at high velocities. They go around it, they can go through it... and they will be able to chase the player around environmental objects as well.

    I believe this is the first time in video game / space sim history that this is possible with 200+ ships running at once at 4 ms response time.

    With 200 ships im pushing 260 to 300 FPS, 4MS response time. Completely stable with not a single frame drop.

    I just cannot wait to show you guys in the next video. The video after that i will post a recording of the editor with my visual debugger on so you can gawk and go WTF is that.. what the... is that.

    Don't ask for a tutorial, the main algorithm has over 240 variables. So yah.. about that tutorial lol....
     
  35. FirstTimeCreator

    FirstTimeCreator

    Joined:
    Sep 28, 2016
    Posts:
    404
    But first before i make this video :p

    I'm going to setup a star and install a hyperspace jump system which will make the ships fly at a velocity of 100,000.

    Currently there is a hyper jump system for in solar system travel that caps off at a velocity of 45,000.

    who is wondering if unity can handle this lol, i guess we are about to find out if you can travel at the speed of light in a game engine (figuratively speaking) 100K is pretty fast for a RB to moving at lol.