Biologically-based learning in the ARBIB autonomous robot

Abstract

We describe the autonomous robot ARBIB, which uses biologically-motivated forms of learning to adapt to its environment. The "nervous system" of ARBIB has a nonhomogeneous population of spiking neurons, and uses both nonassociative (habituation, sensitization) and associative (classical conditioning) forms of learning to modify pre-existing ("hard-wired") reflexes. As a result of interaction with its environment, interesting and "intelligent" light-seeking and collision-avoidance behaviors emerge which were not pre-programmed into the robot-or "animat". These behaviors are similar to those described by other workers who have generally used behaviorally-motivated reinforcement learning rather than biologically-based associative learning. The complexity of observed behavior is remarkable given the extreme simplicity of ARBIB's "nervous system", having just 33 neurons. It does not even have a brain! We take this to indicate that great potential exists to explore further "the animat path to AI".