Unsupervised learning of reflexive and action-based affordances to model navigational behavior

Here we model animat navigation in a real world env ironment by using place cell as a sensory representation. The cells’ place fields divided the environment into discrete states. The robot learns knowledge of the environment by memorizing t he sensory outcome of its motor actions. This was composed of a central process, le arning the probability of state-to-state transitions by motor actions and a distal processin g routine, learning the extent to which these state-to-state transitions are caused by sensory-dr iven eflex behavior (obstacle avoidance). Navigational decision making integrates central and distal learned environmental knowledge to select an action that leads to a goal state. Dif ferentiating distal and central processing increases the behavioral accuracy of the selected a tions. We claim that the system can easily be expanded to model other behaviors, using alterna tive definitions of states and actions.