Recently, I am trying to build a traffic scenario using Sumo and combine it with reinforcement learning algorithms for decision making in the field of autonomous driving (For example, when a red light appears at a intersection, one should stop, or when driving straight at the intersection and seeing a vehicle on the right, one should brake urgently). At present, I have learned that both FLOW and gym frameworks can serve as a “bridge” between simulation scenarios and reinforcement learning. Therefore, I would like to ask which of these two frameworks is better for compatibility and ease of use?

I hope the compatibility performance of this framework can be as good as possible, because after completing this step, I also hope to build a C-V2X communication network for my traffic scenario and build a federated transfer learning framework.