A multi-path selecting navigation framework with human supervision

Abstract

Robotic navigation remains one of the fundamental problems of mobile robotics, especially when some uncertain variables need to be considered. The cost function of the path planning algorithm cannot always represent the optimum path of the task completely in some multi-choice environments. In this paper, an interaction framework was designed for multiple path selection, supervised by an operator with an intuitive user interface. An interaction approach based on inquiry message was proposed to release the workload impact on human operator. A path priority function was presented to guarantee the multiple path selection. Simulation and experiments on Pioneer 3 AT robot showed good performance of our multi-path selecting navigation system when the operator's experiential cogitation need to be considered in path selection. © 2012 Springer-Verlag.