INDOOR LOCALIZATION SYSTEMATICAL ERRORS ANALYSIS FOR AGVs

摘要

Automated guided vehicles (AGVs) are commonly encountered in modern factories for handling materials. This paper studies and compares performance of dead-reckoning, heading measurement and triangulation as a means of locating AGVs in indoors quasi-structured environments. Because global positioning system (GPS) does not work indoors, these other approaches of locating AGVs are adopted. The main result of this paper stems from the dead-reckoning algorithm with accurate heading measurement condition. If the heading is accurately measured, then the ellipse that quantifies measured position error becomes a line with slope equal to the measured heading. This result makes triangulation more efficient for the position error reset procedure than for the dead-reckoning based almost only in rotary encoders. In this case, only one beacon is sufficient to accurately reset position error when the triangulation procedure is carried out. The triangulation hardware is based on the AGV communications sub-system and its feasibility is shown through an experimental result.