Robot Ants - Video

Reinforcement learning demo made with Unity Machine Learning Agents.
The ants want to explore the terrain but also need to find power-ups and recharge. They leave trails proportional to their energy level. Each ant scans its vicinity for trails to follow. Pathways between power-ups emerge over time.

The project utilizes the now outdated version 0.6 of ml-agents. In order to run the trained models, you need to install the Unity TensorFlowSharp Plugin and add ENABLE_TENSORFLOW to Edit > Project Settings > Player > Configuration > Scripting Define Symbols in the Unity Editor. An archived copy of the plugin is available here.

Ant Design: Ergin3D (modified) via turbosquid.