Hello everybody, nice to have you here! This is the place where we will talk about hot topics in developing artificial intelligence for simulations and games. From now on, we will inform you what we are currently doing and what we are going to do next (thanks to you and, of course, with the help of the money you have spent on our products).
How rude of us, we have not even introduced ourselves. So here we go: We are Polarith, a young start-up and full of creative ideas on how to make the lives of artificial intelligence developers much easier. Besides our passion for AI, we love to work with 3D engines so that we decided to found a company and to turn the results of years of AI research into decent Unity plugins you can easily use for your own projects.
Release of Polarith AI
Before a month we released our first plugin: Polarith AI for Movement. It utilizes brand-new context steering algorithms to provide you very powerful tools for creating natural moving agents like you never did before. Once you have learned the basics, it can be a hell of a time saver for simulating autonomous individuals as well as swarms of characters. This is due to the way it balances and combines environmental information which works much more natural and better than anything what classic steering approaches can do, whereby it dramatically reduces the amount of code needed to handle unwanted special cases in movement AI logic. We designed the system in a way which allows you to achieve several different effects just by blending AI behaviours together as easy as Photoshop layers.
These kind of algorithms are so robust that they were used to create the artificial intelligence for the game F1 which got only good critiques for its character movement. Therefore, it was crucial for the developers to make the AI behave as realistic as possible. With the release of our first plugin, we bought the power of these techniques right into Unity at a relatively low cost, because we want that everybody is able to profit by this amazing approach.
Progress since Release
Since the release of our plugin, we have worked hard to improve and polish the documentation of Polarith AI, this includes the content as well as the overall look and feel. Now, there are four major sections full of impressions on how to make the best out of our plugin.
Besides that, we have been working at the first patch for fixing bugs. We are sure that we manage to bring it live during the next month. Rumor has it that there might be a few new but handy features in it as well.
Work in Progress and Future Plans
Now we have released the plugin, we focus more on writing tutorials and publishing videos on how to build a really clever AI for certain situations and game genres.
Moreover, we will keep our promise and make our inbuilt AI behaviours open-source one after another. This also includes the base classes, the back-end and the front-end for (radius) steering behaviours so that you get a better understanding for the system, which allows you to learn more easily how to write your own behaviours.
Unfortunately, we currently do not provide a place for you to ask general questions and discuss interesting things. This is what we are going to do this or the next week. We will be active in the Unity Forum as well as in Unity Answers. On top of that, we plan to create our own support forum in the future, too.
Our focus for the first release lay on the functionality of context steering in general. We know that the overall usability and UI design of the components can be improved though. That is exactly what we are concentrating on next.
At last but not least, here are some hot development news: We are currently working on an own path structure. We will use this to provide you new and impressive inbuilt AI behaviours so that you can easily create things like patrols. Simultaneously, it will be the basis for our own integrated pathfinding in the future. This way we plan to combine the best of both worlds: Local decision making (what steering actually does) and global decision making (the purpose of pathfinding). The result will be an AI behaviour like it was never seen before. It will even work with labyrinth-like environments using context steering, keeping computation costs low most of the time, but using pathfinding whenever necessary to find a proper global solution. We are very excited about what the future brings.