To show some progress, here are some examples of the current state of the project.
What works and what you can see in the example images:
- Map Scanning and persisting to DB
- basic Cluster Generation (Convex Hull) using ML k-means-nearest-neighbour DBSCAN algorithm
- basic Cluster Rendering
- Configuration System (DB based)
- Cluster Entity-Type Configuration
To create more intelligence from gameMap data the following features are planned:
- Visual Map Generation
- Forest detection / forest heatmap
- Structure detection
- Terrain heightmap
- Map shading / lightmap
- Contour lines
- Visualization of slope map (will be used for pathfinding and movement)
- Basic Idea of rasterized "AI" map vision
- Mil detection
- Industrial detection
- Qualification and prioritization of clusters / targets
After having more intelligence from the gameMap data, the work on the AI Commander will start. Planned is to harness the power of OpenAI to create a commander that make decisions and gives orders to his units.
Run com.recom.RecomBackendApplication with "local" profile. application-local.properties is preconfigured to work with the provided docker-mariaDb instance.