Combining associative learning and nonassociative learning to achieve robust reactive navigation in mobile robots

We present a neural network that learns to control approach and avoidance behaviors in a mobile robot based on three forms of animal learning: classical conditioning, operant conditioning, and habituation. Mechanisms of classical conditioning are used to learn to predict the proximity of obstacles and sources of light. At the same time the robot learns through operant conditioning to generate the desired avoidance and approach behaviors. Learning takes place as the robot moves around an environment cluttered with obstacles and sources of light. The neural network requires no knowledge of the geometry of the robot or the configuration of the sensors. After learning the robot can choose among different behaviors depending on the moment-by-moment combination of sensory information and internal needs. We discuss the problem of the oscillatory movements observed when the robot navigates through narrow hallways. Results show that habituation to the proximity of the walls can lead to smoother navigation. Habituation to sensory stimulation to the sides of the robot does not interfere with the robot’s ability to turn at dead ends and to avoid obstacles outside the hallway. This work shows that several biological mechanisms of learning can be combined to produce adaptive behaviors in real mobile robots.