Autonomous locomotion and navigation of anthropomorphic robot

Abstract

The paper describes a software implementation of the locomotion, autonomous navigation, and voice interface support subsystems for the anthropomorphic robot AR-600E. All these subsystems comprise the integer system which provides the capability for autonomous movement of the biped AR-600E robot. The general architecture of the autonomous movement system, as well as each subsystem features, are presented; much attention is paid to computer vision role in navigation and next step planning. The main specificity of the locomotion support system is the use of FRUND modeling CAE system as a high-level movement controller, conditioning some specific demands to route planning. The performance of the robot onboard computer aimed to support all these subsystems is also assessed. The optimization potential of these subsystems to migrate them to the onboard standalone or augmented with the embedded graphics accelerator computer is revealed. Further directions of the investigations are outlined.