This project is a demonstration of artificial intelligence and my personal take on driving AI. This game/environment/demo consists of a player car (the only blue car) in addition to CPU cars. The goal of this entire game is to make a left turn through the intersection losing as little points as possible. Everything is based on points, and doing anything that goes against proper driving will take off points. Using neural networks, the user can choose the amount of iterations to train the AI before it goes through the intersection, and based on that the performance of it will vary. The user can also choose Free Play mode, which will give the controls over from the AI to the user, and using WASD they can try out the environment themselves.
A demo and further explanations can be found here
Make sure you unzip the file before trying to run the program, otherwise you will just see a gray screen. After you have unzipped "driver.zip", run the "driver.jar" from the unzipped folder. It's important to run the "driver.jar" file in the same directory as the "res" folder.
Java and supporting Java libraries. Photoshop to create the assets
Me (Kyle Wu)
Main menu:
- Start button: Start the AI training and demo
- Left arrow: Decrease the training iteration count
- Box with numbers: Iterations
- Right arrow: Increase the training iteration count
- Check mark beside start button: Disable automatically increase the iteration by 5 after every attempt
- Freeplay: User takes control.
Use WASD to drive around for Freeplay mode. In Freeplay mode, when users successfully make it through the left turn and drive off the top of the screen, it will restart automatically after a few seconds.