The mobile robot surrogate method for developing autonomy

Abstract

The mobile robot surrogate (MRS) method, as a means for developing autonomy in complex systems, is presented. Used to analyze and interpret sonar data for landmark identification and navigation, this method involves: 1) putting a human operator within a telerobotic loop; 2) studying the capabilities and limitations of the operator for different tasks; 3) using protocol analysis to extract the knowledge from the operator into an expert system; 4) substituting the expert system for the operator within the robotic loop. A central aspect of the MRS methodology is that the operator, through the telerobotic interface, is limited to the same sensor information as the robot. The strength of this methodology is shown through the development of door identification and door traversal performances by an autonomous mobile robot using ultrasonic data.