Flatland: a Lightweight First-Person 2-D Environment for Reinforcement Learning
This provides a simpler, customizable benchmark for researchers in reinforcement learning, particularly for problems related to lifelong learning, though it is incremental compared to existing 3D platforms.
The authors tackled the need for a lightweight environment for fast prototyping and testing of reinforcement learning agents by introducing Flatland, a 2-D first-person environment that emulates real-world physical properties, and showed that baseline agents can rapidly solve a navigation task in it.
Flatland is a simple, lightweight environment for fast prototyping and testing of reinforcement learning agents. It is of lower complexity compared to similar 3D platforms (e.g. DeepMind Lab or VizDoom), but emulates physical properties of the real world, such as continuity, multi-modal partially-observable states with first-person view and coherent physics. We propose to use it as an intermediary benchmark for problems related to Lifelong Learning. Flatland is highly customizable and offers a wide range of task difficulty to extensively evaluate the properties of artificial agents. We experiment with three reinforcement learning baseline agents and show that they can rapidly solve a navigation task in Flatland. A video of an agent acting in Flatland is available here: https://youtu.be/I5y6Y2ZypdA.