I've been making AI for my game for over a month. It was a long and difficult process, I even lost several kilograms. The result is on the GIF. In my opinion it turned out very well, although not perfect

I am developing my own game engine. Currently, I am adding a graphical user interface. Later, I plan to create games and applications using this engine. The screenshot shows two buttons and a text input field with their colliders (red rectangles)

Now it's very easy to create an animation from a sprite sheet (texture atlas) in my engine! Image credits: - @MatheusBarbiot_ - @Raid_BugRepellent - @ZenOfficial

So I made a smart aiming of the tank turret at a moving target. This required a long mathematical formula, but it did not affect the optimization (I used 1500 tanks on one screen and there were no lags!)

Now the tank turret rotates realistically - at different speeds, depending on the movement of the tank. I tried to do this before, but it turned out very poorly and strange things happened

(Tank turning speed increased by ~4 times)

Rotation of two different types of tanks.

Next, I should improve their aiming, since at the moment tanks cannot hit a moving targets. Well, I think it won't be difficult to make smarter aiming. I hope

First AI tests

At the moment, bots are like zombies - they follow the player and can avoid obstacles in their path, but they don’t always succeed

They are very stupid, but I will try to make them smart

The story of Daswaget. Episode 1: Prototype (or the story of my game engine and diploma project) The screenshot shows my diploma project running on the emulator

I decided to do a tank race to test the AI. Participants: 4 bots and me (gray tank)

I improved my old tank control code. It used to be 400+ lines, but now it's only 66. Only 66, Carl!

This is very cool, but while I was rewriting my code, many bugs appeared, for example, when I clicked on WASD, I controlled all the tanks at once (GIF)