One robot reasons about itself, many robots reason about each other
In the first two courses in this track, a single robot only had to answer two questions: where am I, and how do I get to my goal without hitting something. Every sensor reading, every planned path, every obstacle check was about that one robot and its environment. The moment you add a second robot into the same space, a third question appears that didn't exist before: what is every other robot doing, and how does that change what I should do. That question doesn't scale away, it multiplies. With ten robots sharing a warehouse floor, each one potentially needs to reason about nine others, and the space of possible interactions grows fast.
This is why multi-robot systems are not just 'single-robot navigation, run N times.' A single robot's path planner treats other robots, at best, as moving obstacles to avoid. That's a reactive, last-resort kind of safety, not coordination. Real multi-robot systems need to actively avoid colliding with each other, avoid wastefully duplicating effort like two robots driving to the same shelf, and often need to cooperate to do things no single robot could do alone, like two robots jointly carrying a load too heavy for one.