Continuously estimating an agent's state space and a representation of its
surroundings has proven vital towards full autonomy. A shared common ground
among systems which successfully achieve this feat is the integration of
previously encountered observations into the current state bei