SONAR-BASED MOBILE ROBOT NAVIGATION THROUGH SUPERVISED LEARNING ON A NEURAL-NET

For mobile robot navigation in an unknown and changing environment, a reactive approach is both simple to implement and fast in response. A neural net can be trained to exhibit such a behaviour. The advantage i s that, it relates the desired motion directly to the sensor inputs, o bviating the need of modeling and planning. In this work, a feedforwar d neural net is trained to output reactive motion in response to ultra sonic range inputs, with data generated artificially on the computer s creen. We develop input and output representations appropriate to this problem. A purely reactive robot, being totally insensitive to contex t, often gets trapped in oscillations in front of a wide object. To ov ercome this problem, we introduce a notion of memory into the net by i ncluding context units at the input layer. We describe the mode of tra ining for such a net and present simulated runs of a point robot under the guidance of the trained net in various situations. We also train a neural net for the navigation of a mobile robot with a finite turnin g radius. The results of the numerous test runs of the mobile robot un der the control of the trained neural net in simulation as well as in experiments carried out in the laboratory, are reported in this paper.