Constraining policies to select behavior modules allows reinforcement learning transfer from simulation to real robot swarms.