It's already possible for robots to navigate without maps, but having them navigate well is another matter. You don't want them to waste time backtracking, let alone fall down if they bump into an unexpected obstacle. Facebook might have a solution. It recently developed a distributed reinforcement learning algorithm that not only reaches its destination 99.9 percent of the time without using maps, but can do so with just a three percent deviation from the ideal path. DD-PPO (Decentrialized Distributed Proximal Policy Optimization), as it's called, doesn't need more than a standard RGB camera with depth data, GPS and a compass.