Robot navigation as hierarchical active inference

Localization and mapping has been a long standing area of research, both in neuroscience, to understand how mammals navigate their environment, as well robotics, enable autonomous mobile robots. In this paper, we treat navigation inferring actions that minimize (expected) variational free energy under hierarchical generative model. We find familiar concepts like perception, path integration, localization naturally emerge from active inference formulation. Moreover, show model is consistent with models hippocampal functions, can be implemented silico on real-world robot. Our experiments illustrate robot equipped our able generate topologically maps, correct behaviour inferred when goal location provided the system.