A supervised deep neural network that learns to drive in video games. The main objective of this project is to achieve a model that can drive in Grand Theft Auto V. The model is not expected to drive following traffic regulations, but imitate how humans drive in this game: Drive at full speed through the city avoiding other cars and occasionally humans and lampposts. A marker will be set in the game map, the model should be able to arrive to the marker driving trough the city.
The model is trained using human labeled data. We record the game and key inputs of humans while the play the game, this data is used to train the model.
While we focus on self-driving cars and the video game Grand Theft Auto V this model can be adapted to play any existing video game.
Architecture
Software and HOW-TO
Requirements
Generate dataset
Train the model
Run the model
Authors
github地址：