Homeostatic Control For A Mobile Robot: Dynamic Replanning In Hazardous Environments

Abstract

A longstanding goal of robotics has been to introduce intelligent machines into environments that are dangerous to humans. These environments also pose hazards to the robots themselves. By embedding sensing devices as a means for monitoring the internal state of the robot, dynamic plan reformulation can occur in situations that threaten the existence of the robot. A method exploiting an analogy to the endocrine control system is forwarded as the preferred method for homeostatic control - the maintenance of a safe internal environment for the machine. Examples are given describing the impact of fuel reserve depletion and global temperature stress. A methodology using signal schemas as a means to supplement the existing motor schema control found in the Autonomous Robot Architecture (AuRA) is presented.