Real-time analysis of the robustness of the navigation strategy of a visually guided mobile robot

Abstract

This paper describes a method that is able to analyze in real-time the robustness of a pattern-driven guidance strategy for mobile robots. The robot computes its next step towards the goal by considering visual patterns which have been previously selected through a biologically-inspired phase. Starting from the analysis of the navigation vector field the system generates in the environment, formal aspects related to the concept of convergence can be considered. In this paper we show how the convergence analysis can be effectively addressed in real time, though not in the whole environment but along the path the robot has traveled. A Nomad200 is the agent used for the experiments.