Development of navigation system for autonomous guided vehicle localization with measurement uncertainties

Abstract

Recently, Autonomous Ground Vehicles (AGV) have been dramatically developed in various engineering applications, such as Industry 4.0 manufacturing and smart technology. Mapping navigation plays a critical role in the movement of the AGV in a cluttered environment. Hence, several problems related to this field must be addressed for a wide application of AGV in reality. In this paper, an innovative methodology is proposed for AGV localization with measurement uncertainties. Overall, this approach has a total of three key steps. To begin with, a path planning is designed to establish a safe, effective, and optimal path. Particularly, the visibility graph is built by determining a geometric free configuration space of the AGV. The Dijkstra's algorithm is applied to the visibility graph to find a feasible path which has random starting and ending points. After that, the enhanced triangular decomposition method is utilized to quickly localize the AGV in two-dimensional space. Finally, the navigation system is developed to optimize the pathways for the continuous movement of the AGVs. Extensive experiments are conducted among different scenarios to evaluate the precision and stability of the proposed method.