# GTA V Autonomous Racing Bot

Behavioral cloning AI for autonomous racing in GTA V. Different models to come.

### Project Structure
```
gtav-racebot/
├── src/              # C++ source code
├── include/          # Header files
├── dependencies/     # ScriptHookV + ONNX Runtime
├── build/            # Visual Studio project files
├── bin/              # Build outputs
├── scripts/          # Deployment utilities
└── docs/             # Documentation
```

### Setup Dependencies

**ScriptHookV:**
1. Download SDK from [dev-c.com](http://www.dev-c.com/gtav/scripthookv/)
2. Extract to `dependencies/scripthookv/`
3. Ensure structure:
```
   dependencies/scripthookv/
   ├── inc/
   │   ├── natives.h
   │   ├── types.h
   │   └── ...
   └── lib/
       └── ScriptHookV.lib
```

**ONNX Runtime:**
1. Download [onnxruntime-win-x64-1.16.3.zip](https://github.com/microsoft/onnxruntime/releases)
2. Extract to `dependencies/onnxruntime/`
3. Ensure structure:
```
   dependencies/onnxruntime/
   ├── include/
   │   └── onnxruntime_cxx_api.h
   └── lib/
       ├── onnxruntime.lib
       └── onnxruntime.dll
```

**Afterwards, must set up configs in the settings in the .vcxproj file!**

### Build
1. Open `build/TEMPLATE.sln` in Visual Studio
2. Build → Rebuild Solution (Release, x64)
3. From the build folder, copy and paste the .asi file to GTAV.exe.

### Usage
1. Launch GTA V (Story Mode)
2. Get in the vehicle
3. Press **F10** for AI mode

### Features
- Real-time inference with ONNX Runtime
- 7-ray LiDAR sensor simulation
- In-game data logging
- Debug visualization

### Tech Stack
- C++ (Visual Studio 2022)
- ScriptHookV
- ONNX Runtime 1.16.3
- Python (Keras/TensorFlow for training)